These materials are © 2014 John Wiley & Sons, Inc.

Any dissemination, distribution, or unauthorized use is strictly prohibited.

 HP NonStop

comForte Special Edition

by Werner Alexi,
Thomas Burg, and
Bill Sempf

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 HP NonStop For Dummies®, comForte Special Edition
Published by
John Wiley & Sons, Inc.
111 River St.
Hoboken, NJ 07030-5774
www.wiley.com
Copyright © 2014 by John Wiley & Sons, Inc., Hoboken, New Jersey
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form
or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as
permitted under Sections 107 or 108 of the 1976 United States Copyright Act, without the prior written
permission of the Publisher. Requests to the Publisher for permission should be addressed to the
Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011,
fax (201) 748-6008, or online at http://www.wiley.com/go/permissions.
Trademarks: Wiley, For Dummies, the Dummies Man logo, The Dummies Way, Dummies.com, Making
Everything Easier, and related trade dress are trademarks or registered trademarks of John Wiley &
Sons, Inc. and/or its affiliates in the United States and other countries, and may not be used without
written permission. comForte and the comForte logo are trademarks or registered trademarks of
comForte 21 GmbH, Germany. HP NonStop is a trademark of Hewlett-Packard Development
Company, L.P. All other trademarks are the property of their respective owners. John Wiley & Sons,
Inc., is not associated with any product or vendor mentioned in this book.

LIMIT OF LIABILITY/DISCLAIMER OF WARRANTY: THE PUBLISHER AND THE AUTHOR MAKE

NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE ACCURACY OR COMPLETE-
NESS OF THE CONTENTS OF THIS WORK AND SPECIFICALLY DISCLAIM ALL WARRANTIES,
INCLUDING WITHOUT LIMITATION WARRANTIES OF FITNESS FOR A PARTICULAR PURPOSE.
NO WARRANTY MAY BE CREATED OR EXTENDED BY SALES OR PROMOTIONAL MATERIALS.
THE ADVICE AND STRATEGIES CONTAINED HEREIN MAY NOT BE SUITABLE FOR EVERY SITU-
ATION. THIS WORK IS SOLD WITH THE UNDERSTANDING THAT THE PUBLISHER IS NOT
ENGAGED IN RENDERING LEGAL, ACCOUNTING, OR OTHER PROFESSIONAL SERVICES. IF PRO-
FESSIONAL ASSISTANCE IS REQUIRED, THE SERVICES OF A COMPETENT PROFESSIONAL
PERSON SHOULD BE SOUGHT. NEITHER THE PUBLISHER NOR THE AUTHOR SHALL BE LIABLE
FOR DAMAGES ARISING HEREFROM. THE FACT THAT AN ORGANIZATION OR WEBSITE IS
REFERRED TO IN THIS WORK AS A CITATION AND/OR A POTENTIAL SOURCE OF FURTHER
INFORMATION DOES NOT MEAN THAT THE AUTHOR OR THE PUBLISHER ENDORSES THE
INFORMATION THE ORGANIZATION OR WEBSITE MAY PROVIDE OR RECOMMENDATIONS IT
MAY MAKE. FURTHER, READERS SHOULD BE AWARE THAT INTERNET WEBSITES LISTED IN
THIS WORK MAY HAVE CHANGED OR DISAPPEARED BETWEEN WHEN THIS WORK WAS WRIT-
TEN AND WHEN IT IS READ.

For general information on our other products and services, or how to create a custom For Dummies
book for your business or organization, please contact our Business Development Department in the
U.S. at 877-409-4177, contact info@dummies.biz, or visit www.wiley.com/go/custompub. For
information about licensing the For Dummies brand for products or services, contact
BrandedRights&Licenses@Wiley.com.
ISBN 978-1-118-84993-4 (pbk); ISBN 978-1-118-85012-1 (ebk)
Manufactured in the United States of America
10 9 8 7 6 5 4 3 2 1

Publisher’s Acknowledgments
Some of the people who helped bring this book to market include the following:
Project Editor: Jennifer Bingham
Acquisitions Editor: Katie Mohr
Editorial Manager: Rev Mengle
Business Development Representative: Sue Blessing
Project Coordinator: Melissa Cossell

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Table of Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
About This Book......................................................................... 1
How This Book Is Organized..................................................... 2
Icons Used in This Book............................................................. 2
Beyond the Book......................................................................... 2

Chapter 1: What Is HP NonStop? . . . . . . . . . . . . . . . . . . . . 3

Why the World Needs HP NonStop.......................................... 3
The increase of data......................................................... 3
Fault tolerance and high availability explained............ 4
Data integrity, reliability, and RAS................................. 5
Why HP NonStop Runs Nonstop............................................... 6
Fault-tolerant hardware................................................... 6
Fault-tolerant software..................................................... 6

Chapter 2: A Platform Overview. . . . . . . . . . . . . . . . . . . . . 7

Some Core Concepts.................................................................. 7
Processes within HP NonStop......................................... 7
You have mail: Messages............................................... 10
How processes and messages work together............. 11
A self-managing and self-healing cluster in a box....... 11
Layers or Personalities of the Operating System................. 12
NonStop OS is the name of the operating system...... 13
OSS is a Unix-like personality........................................ 13
Guardian is the lower layer........................................... 14
Working with Files in the Guardian Environment................. 15
Putting the Data in a Base........................................................ 16
TMF: Keeping track of transactions............................. 17
Enscribe is the original database................................. 18
SQL/MP brought structured queries............................ 19
SQL/MX is the gold standard........................................ 19
Pathway – the Transaction Monitor....................................... 20
Running Two HP NonStop Systems in
Active-Active Mode............................................................... 21
Key Benefits of HP NonStop Systems..................................... 22
Common standards, uncommon advantages.............. 22
Low TCO (total cost of ownership).............................. 23
Excellent support............................................................ 23
Cloud/hybrid computing............................................... 24
HP NonStop on X86........................................................ 24

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 iv HP NonStop For Dummies, comForte Special Edition

Chapter 3: Developing Software for HP NonStop . . . . . 25

Developing for HP NonStop Doesn’t Mean
Developing on HP NonStop.................................................. 25
Programming languages and development
environments.............................................................. 26
Using Eclipse to develop for HP NonStop................... 26
Enter Java........................................................................ 28
Debugging is in the mix.................................................. 28
Writing Good Software for HP NonStop................................. 29
Assembling a team of developers................................. 29
Writing efficient code..................................................... 29
Nonstop programming................................................... 30
No-waited programming................................................ 30
Modernizing Existing Applications on HP NonStop............. 30
Porting Existing Applications to HP NonStop....................... 31
Database migration........................................................ 31
Application migration.................................................... 32
Further considerations.................................................. 32

Chapter 4: HP NonStop as Par t of Enterprise IT . . . . . . 33

Securing HP NonStop Systems................................................ 33
Security has come a long way....................................... 34
General platform hardening.......................................... 34
Specific hardening of the HP NonStop platform......... 35
IAM: Identity and access management........................ 35
Keys and certificates...................................................... 36
Integrating HP NonStop with Other Systems........................ 36
Client-server and multitier design................................ 37
Raw messaging................................................................ 37
Simple Object Access Protocol (SOAP)....................... 37
Representational State Transfer (REST)...................... 38
Operations Management.......................................................... 38
File Transfer............................................................................... 39
Messaging.................................................................................. 39

Chapter 5: Ten Resources for More Information. . . . . . 41

NonStop Terminology and Entities: A Quick Reference...... 41
TACL Fact Sheet........................................................................ 41
Open Source Tools Available on HP NonStop....................... 42
Tips for Developing for HP NonStop...................................... 42
No-waited Programming.......................................................... 42
Modernizing Legacy Applications.......................................... 42
Securing NonStop Systems...................................................... 43
The NonStop ISV Ecosystem................................................... 43
The HP Website and HP NonStop Computing....................... 44
The User Community................................................................ 44
These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Introduction
S ixteen seconds after an outage of a central computer
affecting ATM availability, the event was reported on
Twitter. Minutes later, a storm of Twitter messages appeared.
And a short time after that, the bank’s CEO got a call from one
of the country’s leading newspapers.

Due to a computer outage lasting only a few hours, a major

company in the financial industry lost an estimated amount of
US$90 million.

Are events like that inevitable like thunderstorms or earth-

quakes? Insiders know that they’re not: The HP NonStop Server
can host applications that run for months or years — without
any interruption and without losing a single transaction.

About This Book

This book is for anyone with an interest in HP NonStop sys-
tems. This book is for you if you’re new to the platform, have
spent many years working with it, if you manage people or IT
systems, or if you’re a system operator or developer on/for
the HP NonStop platform.

This is obviously a short book. There isn’t a lot of in-depth

technical information here. This book will help formulate a
strategy for best making use of HP NonStop in a contempo-
rary enterprise environment.

It will not help you design an effective, fault-tolerant nonstop

process pair. But it will describe what one is.

This book was written with and for comForte.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 2 HP NonStop For Dummies, comForte Special Edition

How This Book Is Organized

This book is structured into the following chapters:

✓ Chapter 1: What is HP NonStop? Get started with HP

NonStop. Learn what it does well and how it can help you.
✓ Chapter 2: A Platform Overview: This chapter explains
what makes HP NonStop special and how HP NonStop com-
puting is baked right into the hardware and software layers.
✓ Chapter 3: Developing Software for HP NonStop: Nice
server you got there. Who’s going to code for it? You will
find those answers here.
✓ Chapter 4: HP NonStop as Part of Enterprise IT:
Whether you’ve been running the same application for
20 years or just started afresh, you will want to get HP
NonStop playing nicely with all your other servers. Also,
you want to make sure the platform is secured properly.
✓ Chapter 5: Ten Resources for More Information: We
couldn’t fit everything we wanted into this book, so
­comForte has created some additional content. This
chapter includes links to that content.

Icons Used in This Book

Sometimes, important details are called out using icons.
These are cool little graphics on the left margin of the page.
Keep an eye out for them!

Tips are particularly interesting items in the text that offer

you some practical information you can use.

The Remember icon points you to things that are worth

remembering.

Beyond the Book
For more information on HP NonStop and comForte, go to
www.comforte.com.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 1

What Is HP NonStop?
In This Chapter
▶ Examining real-world uses for reliable computing
▶ Looking at the components of HP NonStop

H P NonStop is a set of hardware and software that has

been designed and built to survive the loss of compo-
nents that most other systems would consider fatal. It is self-
healing, redundant, and reconfigures itself in case of failure.
HP NonStop is used by organizations that need computers
that never stop working — thus the name.

Why the World Needs

HP NonStop
Wherever you see a system that simply can’t be allowed to
fail — under any circumstances — you will probably find an
HP NonStop server running the show.

Telecommunications, automated teller machines (ATMs), emer-

gency phone services, point of sale (POS) systems, and manu-
facturing facilities all depend on HP NonStop to reduce global
downtime to just a few — or even zero — minutes per year.

The increase of data

The Internet has created a dramatic increase in the amount of
data being processed and stored. Add new or newish trends
such as Big Data and the Internet of Things and you end up
with even more data.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 4 HP NonStop For Dummies, comForte Special Edition

Some data has the two additional characteristics

✓ Must be available all the time

✓ Must be consistent at any point in time

Operations on that type of data are called transactions. For

instance, it’s not okay for your ATM to take the money from
your checking account and not give it to you. Either the
system performs all the steps required to complete the trans-
action or none of them. Similarly, it isn’t okay to tell someone
that the police are on the way to the scene of the accident and
lose the message that dispatches the police. Money (lots of!),
lives, and reputations can be lost by a single message (trans-
action) getting lost.

Transactions are the lifeblood of corporations.

Fault tolerance and high

availability explained
Fault tolerance is the capability of a system to continue to
function in a predictable fashion as components fail. HP
NonStop systems are designed to be fault tolerant; therefore,
no single point of failure will impact the system and, in fact,
even many components failing will most likely not impact
another one.

Systems that can’t be reached can’t be used — being unavail-

able means losing money. High availability, as defined by
International Data Corporation (which defines such things),
means an application is available even in the face of failure in
part of the system. High availability means avoiding any out-
ages, whether planned or unplanned. Systems are rated from
Availability Level 1 to 4. HP NonStop systems use a combina-
tion of hardware and software to achieve the highest level
of availability (technically termed Continuous Availability or
Availability Level 4). Level 4 is often described as five 9s or
higher, meaning an uptime of 99.999 percent or more, which
translates into a total downtime of less than five minutes per
year at five 9s or about 30 seconds at six 9s.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 1: What Is HP NonStop? 5
Data integrity, reliability,
and RAS
Data integrity is the quality of maintaining data correctly,
despite flaws in surrounding systems. HP NonStop is designed
from the ground up for data integrity — which is a precondi-
tion for reliability. A reliable system will not return a result
rather than returning a wrong result or corrupting a database.

If a system doesn’t have three key properties: reliability, avail-

ability, and serviceability (RAS), it can’t be depended on to
perform mission critical operations.

The original designers of HP NonStop built the platform for fault

tolerance, high availability, data integrity, and RAS. The plat-
form still offers those today and into the years ahead. For more
on the original design of HP NonStop, see the nearby sidebar.

A history of HP NonStop
In 1974, Tandem Computing entered in 1976, beginning a nearly 40-year
the large-scale computing market relationship between the banking
with the revolutionary idea that com- industry and fault-tolerant comput-
puter services should be predictable ing. In 1997, Tandem was acquired by
and dependable. Computing was very Compaq along with competitor DEC.
different in those days: Computers In 2002, Compaq and HP merged,
were powerful, but unreliable. bringing Tandem back under the
company where it got started.
James Treybig, an HP engineer,
assembled some of the best and Although the core concepts of HP
brightest engineers and designed NonStop have remained stable,
a set of hardware and software the HP NonStop system of today
that provided a fault-tolerant online has very little in common with the
transactional processing (OLTP) very first Tandem system of the
environment for business. late 1970s  —  except the almost
unmatched availability.
The first Tandem/16 system rolled off
the line and was installed at CitiBank

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 6 HP NonStop For Dummies, comForte Special Edition

Why HP NonStop Runs Nonstop

Today, a wide range of hardware platforms run an even wider
range of operating systems. And with few notable exceptions, the
vendors of the hardware and software pieces aren’t the same. HP
NonStop is a tightly integrated suite of hardware and software
that gathers much of its strength from this exact tight integration.

Fault-tolerant hardware
Quality servers in the Internet space often come with two
power supplies and a RAID array. RAID, or redundant array
of independent disks, is a simple and effective way to provide
basic fault tolerance to storage.

But even for standard applications, this level of fault tolerance

often isn’t sufficient, so computer vendors invented clusters. A
cluster is a tightly coupled set of multiple systems sharing cer-
tain resources and protecting against certain failures. Clusters
aren’t typically easy to set up and manage, especially when data
has to be kept synchronous between the individual systems.

HP NonStop hardware takes redundancy to the logical extreme,

providing effectively 100 percent uptime. Every hardware com-
ponent in an HP NonStop system is duplicated. At least two
storage units, two RAM busses, and two processors reside in
every HP NonStop system. But that’s only part of the story.

Fault-tolerant software
From an eagle-eye perspective, a single HP NonStop system can
be described as a self-managing, self-healing cluster in a box.

This works by having a rather singular operating system

concept, combined with hardware redundancy: Each CPU
runs independently of each other in a shared nothing environ-
ment, preventing a sick or malicious program from hurting
the system. Critical operating system processes run as what
is called process pairs spanned over two CPUs — this allows
processes to survive failures of the CPU they run in.

Through this design, the software is capable of transparently

failing over to a second set of components without interrup-
tion of the currently executed operations.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2

A Platform Overview
In This Chapter
▶ Explaining core concepts
▶ Examining important components of HP NonStop systems
▶ Looking at benefits of HP NonStop systems

A platform can claim five or six 9s of availability in

theory (see Chapter 1 for the definition of five 9s), but
backing it up in reality is more difficult. The question is: What
about the HP NonStop platform makes it so reliable? This
chapter goes into the answer in some depth.

Some parts of this chapter get rather technical for a For

Dummies book, so if you don’t need to understand every piece of
the magic that makes HP NonStop systems run nonstop, feel free
to gloss over parts of this chapter. We do recommend to at least
read the parts about databases, TMF, and Pathway. We also think
the section “Key Benefits of HP NonStop Systems” is valuable.

Some Core Concepts

At the core of the functionality of HP NonStop is a message-
based architecture. This architecture surrounds the two core
concepts of the HP NonStop operating system: processes and
messages. The following sections discuss these concepts and
also explain why they’re so great.

Processes within HP NonStop

All systems have processes — they’re the basis for contem-
porary computing. A process is any program running on
the system. There are application processes and system

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 8 HP NonStop For Dummies, comForte Special Edition

processes: They perform different functions but they behave

the same.

For example, if you have a Windows machine in front of you,

Task Manager will display the process list for you. Some of
those are applications you’re currently running. But most are
processes that are running without you ever seeing them.

Are you running Linux or Macintosh? Open a command prompt

and type ps -A. You’ll see a similar list of processes — some
of those will be apps you might be running right now, but most
of them are processes running under the covers.

Of course, HP NonStop has the same capability to list pro-

cesses, and there are multiple ways to do it. From the com-
mand shell, typing STATUS * will list all processes (for more
on command-line syntax, see Chapter 5).

Figure 2-1 shows a Windows Explorer style view of the pro-

cesses running on HP NonStop, sorted by the column CPU
busy and showing the busiest processes first. Everything in
the system is implemented as a process: network connectivity,
disks, command shells, transaction monitoring, databases, the
kernel itself — everything.

Processes running in the background doing important work

exist on other systems as well (called services in Windows or
demons in Unix). However, there are two important differences
in the way processes work on HP NonStop:

✓ All critical operating system processes run as process

pairs spanned over two CPUs. Any process that encounters
a fault is automatically switched to its backup instance.
When the backup instance has taken over, after only a few
milliseconds it will then become the primary (and create a
new backup instance). Process pairs implement all central
operating system services. They provide access to the
network, storage, and so on. A single failure of any of these
services will leave the system as a whole intact.
✓ Many processes on HP NonStop have well-known names
(for example, $DATA1). Named processes can exist con-
currently with the same name on two different CPUs.
So even if one CPU fails, the process $DATA1 will still be
there and usable by the application. Usage of a process
means exchanging messages with it.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 9

Figure 2-1: L ooking at processes using the comForte TOP product.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 10 HP NonStop For Dummies, comForte Special Edition

You have mail: Messages

Every operating system is based on the principal of delega-
tion. If you can’t do something that needs to be done, find
someone who can and hand the task over. The basic agents
within systems are processes.

The main difference between HP NonStop and other systems

is the way delegation is implemented. If process A wants pro-
cess B to do something, A puts together a letter describing
what is to be done (a message) and sends it to B. B does its
work and sends a reply back to A. The method to send mes-
sages from A to B is part of the operating system itself. Other
types of operating systems use methods for delegation requir-
ing a much tighter coupling of A and B (like sharing common
memory).

What is so special about using messages?

✓ Location independence: A message from A to B can

travel within the same CPU, to a different CPU, or to a dif-
ferent node in the HP NonStop network, possibly halfway
around the globe. Neither A nor B has to know or care.
If the CPU B used to work on becomes too busy, put a
new one into your system and move B to this new CPU.
Problem solved.
✓ Fault tolerance when using process pairs: Technically,
B can consist of two instances: B1 and B2 running on dif-
ferent CPUs (a process pair). A message from A to B will
only go to one instance — say B1. If B1 (or the CPU where
B1 runs) fails, the operating system can automatically
route the message to B2. B2 does the work and replies.
A doesn’t even notice that there was a fault.
✓ Isolation: If B becomes internally corrupted by a defect
(every program has defects), B can’t corrupt A. The reply
to A’s message might be missing or empty or corrupted,
but A will be able to find out and stay intact. This greatly
improves the overall robustness and security of the system.
✓ Performance manageability: HP NonStop administrators
know their processes and what they’re good for. By mea-
suring resource consumption on a process basis, they
can easily identify performance problems.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 11
There are a lot more advanced advantages of messages (like
easy interceptability or transaction inheritance).

How processes and messages

work together
Every process has its own mailbox, called $RECEIVE, which is
used to read and reply to messages arriving there. The mailbox
maintains a queue of messages, so that the serving process can
work on them one by one. This applies to application programs
as well as services provided by the operating system.

If an application process wants to write something to disk, it

will call a WRITE API function just like on any other operating
system. However, on HP NonStop, the kernel will eventually
translate that API call into a message landing in the $RECEIVE
queue of the proper disk process. The disk process will only
mark this message as completed (by a specific REPLY message)
after the data is safely cached or actually written to disk.

The message-based processing and complete redundancy

complete the picture of a truly fault-tolerant system. If a CPU
fails, or a disk drive, a disk controller, or a network card fails,
HP NonStop systems keep working, moving their messages to
other processes and the work continues.

In Figure 2-1, you can see process $HTTP, which is a simple

web server busily serving static pages and writing log mes-
sages to disk. It has been configured to write a huge amount of
log messages in order to create high disk drive activity. Look
at $data1. This is the process associated to the disk drive
where the log files are being written to. The next process
down the list, called $ZPTM1, is related to the TCP/IP network
stack, which is also implemented as a process, interfacing the
process $http through message passing.

A self-managing and self-healing

cluster in a box
It might seem a lot of effort and maybe even wasteful to imple-
ment a web server by passing messages around between the
actual web server, the disk (process!), and the network (process!)
rather than by moving pointers to buffers through the kernel.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 12 HP NonStop For Dummies, comForte Special Edition

However, it is exactly this design that gives HP NonStop sys-

tems their unique capabilities such as extreme reliability and
scalability — all baked right into the operating system.

All critical processes on HP NonStop are made self-healing by

running as process pairs. Also, all processes can and do run
in different CPUs, even in different systems that are physically
located far apart that communicate via messages.

When you start developing your own application on HP

NonStop, you don’t have to worry about process pairs and the
message system — the development framework will take care
of this for you. See Chapter 3 for more information on this.

Layers or Personalities of
the Operating System
HP NonStop systems are powerful beasts that have evolved
over decades. That resulted in a complex operating system
that has several pieces best understood as layers or personali-
ties. See Figure 2-2.

Figure 2-2: L ayers of the HP NonStop operating system.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 13
NonStop OS is the name
of the operating system
As of today, the name of the unique operating system that
runs on HP NonStop systems is NonStop OS. It consists of sev-
eral components working together.

OSS is a Unix-like personality

From an operator or developer perspective, contemporary
HP NonStop systems can look a lot like Linux or Unix: The
so-called Open System Services (OSS) layer or OSS personal-
ity provides a Unix-like interface to the underlying operating
system kernel. It provides nearly everything that Unix users
are used to having at their fingertips. It has a hierarchal file
system with long and case-sensitive filenames and directories.
OSS also provides most of the Unix utilities, like bash shell,
grep, awk, sed, and many of the programming functions.

It works with the NonStop OS and makes use of many of its

services like fault tolerance and parallel processing. However,
remember that OSS isn’t an operating system, and it doesn’t
replace portions of the HP NonStop OS.

In fact, OSS is so much like Unix that many of the open source
tools and applications that are available in the Unix world are
available for OSS. See Chapter 5 for details.

A core understanding of Guardian (see the next section), the

underlying operating system, is still helpful (especially for
system architects) because OSS is built on top of it. Every OSS
process is a Guardian process and every OSS file is a Guardian
file (and vice versa). There are mapping tables that provide
the long filenames and sensible process identifiers, but under
the covers, the Guardian system is still running the show.

OSS is so powerful that it’s possible to start and end many

development projects completely within the environment.
The Guardian principles are still there, with processes pass-
ing queued messages to other processes’ $RECEIVE inbox.
The development environment, language, and tools, how-
ever, look exactly like Unix.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 14 HP NonStop For Dummies, comForte Special Edition

Guardian is the lower layer

Guardian was the original name of the HP NonStop’s OS. Today,
the term is used to describe an environment that includes:

✓ The operating system kernel itself

✓ The set of APIs specific to the platform, including pro-
cess control, access to files
✓ Tools and utilities that operate on Guardian objects —
files, processes, devices, and so on
✓ Services that run in the Guardian environment

Guardian includes a number of subsystems that are made up

of several communicating processes (no surprise there), each
with its own toolset for management and configuration. To
mention just a few:

✓ The spooler maintains printers and jobs to be printed.

There are HP NonStop configurations that support thou-
sands of printers in nodes all over the globe, and keeping
them all working is the spooler’s job.
✓ The Subsystem Control Facility, or SCF, helps to manage
several subsystems in a persistent fashion. SCF manages
the network, disks, and even the kernel itself.
✓ The Event Management System (EMS) collects and dis-
tributes failure, exception, and information events in a
central place, and distributes them as needed.
✓ MEASURE provides performance metrics on a very
detailed level. It is an important part of the performance
analysis and tuning of an HP NonStop system.

If you find maintenance of the various subsystems in Guardian

a chore, comForte (and other companies) have products that
give a nice graphical user interface for the whole HP NonStop
system.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 15

Working with Files in the

Guardian Environment
Files in the OSS environment are pretty much like files on any
Unix/Linux system. However, files residing in the Guardian
environment differ massively in that there are no directories
or extensions. This lack of structure can be somewhat of a
shock for folks coming from the Windows or Unix world.

At first glance, the Guardian file system seems very limited

and cumbersome. With a little use and using Table 2-1 for ref-
erence, it becomes easy to use and organize.

Filenames in Guardian aren’t case sensitive, and only contain

letters and digits, with a few notable exceptions. Table 2-1
shows how a filename is broken into four parts.

Table 2-1 Parts of a Filename

Part Description Example
Part 1 Name of the node in a network of HP \NS1
(optional) NonStop systems. This always starts with
a backslash (\).
Part 2 Disk name. Starts with a dollar sign ($), $DATA01
(optional) followed by a letter and can only contain
letters and numbers.
Part 3 Subvolume name. This is a lot like a direc- DOCS
(optional) tory. It has to start with a letter and can
only contain letters and digits.
Part 4 Filename. This can be 8 characters. It has ABOOK
(required) to start with a letter and can only contain
letters and digits.

Assuming proper defaults are set (somewhat similar to the

cd command in Windows or Unix), the exact same file from
Table 2-1 can be referenced as ABOOK or DOCS.ABOOK or
$DATA01.DOCS.ABOOK or, finally, \NS1.$DATA01.DOCS.
ABOOK. The last version is the so-called fully qualified filename

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 16 HP NonStop For Dummies, comForte Special Edition

and it is unique across a whole network of HP NonStop sys-

tems. That’s all you get — no extension, no directories.

It should be noted that Part 2 of Table 2-1 can refer to a physi-

cal disk or a virtual disk.

Putting the Data in a Base

Transactional databases are one natural endpoint of all the
fancy concepts of message-based architecture. Although
transactions are supported in IBM’s DB2, Oracle, Microsoft’s
SQL Server, MySQL, and most other databases, the architec-
ture of HP NonStop provides a particular feature that none of
the others have: The transactions and the database are man-
aged by the operating system.

Understanding how to properly process transactions is impor-

tant to understanding the value of HP NonStop. A transac-
tion is a bracket. This bracket allows treating a sequence of
database operations as one single operation from an applica-
tion point of view. The standard example of a transaction is
moving money from Alice’s account to Bob’s. This transaction
must have the following properties:

✓ Atomicity: It must happen completely or not at all —

under no circumstances may the money disappear from
Alice’s account without appearing on Bob’s.
✓ Consistency: At the end of the transaction, the total
amount on Bob’s and Alice’s accounts must be the
same — no money appeared or disappeared.
✓ Isolation: While the money is moving from Alice to Bob,
other transactions on other accounts may be going on
without interfering with the activity on Alice’s and Bob’s
account.
✓ Durability: After the end of the transaction, the new
accounts must be safely stored, so that even switching
the systems off will not undo the action.

The first letters of the words in brackets give the acronym

ACID. ACID is needed whenever transactions carry really
valuable data (like money). ACID is built into the HP NonStop
operation system.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 17
TMF: Keeping track
of transactions
Transaction Monitoring Facility (TMF) is the subsystem of
Guardian that is responsible for managing transactions. Because
TMF is part of the operating system, it provides certain features
that are rarely found elsewhere (if at all).

For each transaction — as defined by the application — TMF

creates a unique identifier, the transid.

✓ The transid can be shared among multiple processes

communicating via the message system to execute a
single transaction. $A starts a transaction, and $B takes
over, using the transid inherited with the message in the
$RECEIVE inbox. This concept allows for some remark-
able modular application design that allows distribution
of work to several processes.
✓ Processes can switch transaction context within the
work cycle, further increasing flexibility in application
design.
✓ TMF provides a common audit trail, providing the basis
for products that allow the near-real-time replication of
database operations to other locations while the system
is running. Also, the audit trail provides a single, atomic
view of all database changes, which allows analyzing
them or rolling back individual changes.
✓ In the case of a total system failure, a consistent state of
the database can be recovered by TMF in a reasonably
short time frame.

This core functionality is baked right into the HP NonStop OS,

but the overlying database management system has changed
with the times.

Currently there are three database systems available under

HP NonStop OS, becoming more and more sophisticated.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 18 HP NonStop For Dummies, comForte Special Edition

Enscribe is the original database

When Enscribe was first written, SQL databases didn’t exist
other than as a concept. Enscribe works on Guardian files.
Most Enscribe files are sets of structured records. The record
structure is described in a data dictionary, and defined using
a custom Data Definition Language (DDL). It works so well, it
still comes as part of every HP NonStop system. Enscribe’s
creators clearly were doing something right.

Enscribe uses four file types:

✓ Key Sequenced Files contain variable length records

found using a key. Keys are data — like customer
number or last name — even combinations of fields.
Key Sequenced Files support insert, update, and delete
operations. They are optionally extended by so-called
Alternate Key Files, which allow fast searches against
multiple different keys in a single file.
✓ Relative Files consist of fixed-length records that are
indexed by a record number. The disk location of the
record can be computed from the record number so no
lookup is required to retrieve data. Inserts, updates, and
deletes are supported. Inserts have to be at the end of
the file or can reclaim space from deleted records.
✓ Entry Sequenced Files are variable length records that
are stored in sequence. Deletes are rare and usually
cause unclaimed space. Updates can be made only if
the data length is the same and are rare as well. Entry
Sequenced Files are often used for log files and such.
✓ Unstructured Files are long strings of bytes, just like a
Unix or Windows file. It is totally up to the application
how the file is searched or updated. No deletes are avail-
able, because there are no structured records.

That’s pretty much it. Enscribe doesn’t provide the features

you might be used to in terms of joins and views and triggers.
Enscribe might be just powerful enough for your purpose, and
it is fast and totally reliable.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 19
SQL/MP brought structured
queries
SQL/MP was the first true relational database system on the
platform. It is still the workhorse of many contemporary
HP NonStop applications. The MP in the name stands for
massively parallel and it isn’t kidding. The principles of the
message-based architecture are applied to the database man-
agement system in such a way that transactional files and
processes are distributed over multiple disks, multiple pro-
cessors, or even multiple systems.

There are still, however, pieces missing from SQL/MP that

would make it a truly relational database system. For instance,
relational integrity isn’t enforced at the data layer — it’s the
application’s problem. Also, there are no triggers or stored
procedures.

All of that said, SQL/MP provides linear scalability so you

get more throughput in a truly linear fashion when you add
CPUs or whole new systems and make them available to your
application.

SQL/MX is the gold standard

SQL/MX is a modern relational database system that was
designed from the ground up to be a reliable, scalable, and
capable database for HP NonStop, making full use of the OSS
personality. Among the three databases, SQL/MX is the prod-
uct that receives the most attention from HP when it comes to
adding new features.

Here are a few of the features of SQL/MX that make it the gold
standard:

✓ In general, SQL/MX provides a highly increased compat-

ibility to the ANSI SQL standard when compared to SQL/
MP. Examples are the capability to use incredibly com-
plex queries or more orthogonality in the query language.
✓ The record size in SQL/MX is 32K versus 4K for SQL/MP.
✓ Stored procedures are available.
✓ Referential integrity is built into the database.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 20 HP NonStop For Dummies, comForte Special Edition

✓ For ODBC connections, precompiled SQL is kept in cache

to prevent costly rebuilds.
✓ You can use SQL/MX from Java, due to fault-tolerant
JDBC drivers.
✓ There is a publish and subscribe mechanism.
✓ There are ANSI standard column and table names.

Pathway – the Transaction
Monitor
HP NonStop systems have always been famous for their linear
scalability. By adding more hardware resources to the system,
you get more transactions per second — in a linear way.

This isn’t true for most other architectures. Typical Unix or

Windows systems will scale to a certain point in a transactional
environment. But from then on, putting more CPUs into the
systems or more systems into the cluster won’t get you more
throughput as you add more hardware; the increase in through-
put will eventually decrease until it hits a final boundary.

Applications making full use of the scalability can be based

on the resource monitor Pathway, which is available from HP
and tightly coupled with the OS. Pathway fully leverages the
core concepts explained earlier in this chapter and provides a
framework to build reliable and scalable applications.

In the original Tandem computers, Pathway would sit in

between the green screen apps written in Screen COBOL and
the HP NonStop processes, determining which process was
the most capable to take on the task requested by the user.
Many of these applications are still in use today, decades later,
and work very well.

Today’s user expects a fancy user interface, or even web

accessible tools. Fortunately, the basic Pathway concept
still works today — better than ever. That is because from
very early on, Pathway provided a clear architectural
separation between a client requesting units of work and
a server delivering them. The most important concept is
that of a serverclass. The term serverclass simply means a

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 21
name for a set of processes, which all can deliver the same
service. Each individual process is called an instance of the
serverclass. A client program — whether it executes locally
on HP NonStop or on a midtier application server or on a PC
client — can send a message to a serverclass. Pathway will
find out the best instance, send the message to that instance,
and reroute the response to the client or, as it was named
predating the terminology of client-server, to the requester.

For a better understanding of the concept consider an anal-

ogy. Imagine a check-in hall at an airport. Each airline can be
compared to a serverclass. So you have a serverclass Delta
and a serverclass Lufthansa. If you have a certain application
task to do, like “Check in for Lufthansa flight LH123,” you find
the queue for the Lufthansa counters. Fortunately, there are
five agents servicing your requests. These five agents all do the
same job — check in passengers for Lufthansa flights. So, the
agents can be compared to the instances of the serverclass.

To take that a bit further, imagining that a big flight comes in

and suddenly the application has to check in 500 passengers all
in a hurry. Scalability can simply be achieved by increasing the
number of agents. This is how massively parallel processing is
achieved.

In the analogy, Pathway would automatically add more agents

(serverclass instances) to the Lufthansa service when the load
increases. The serverclass instances would be naturally dis-
tributed across the CPUs of the system or even across multiple
systems.

Running Two HP NonStop

Systems in Active-Active
Mode
Working together, TMF and Pathway give you the founda-
tions for very high uptime. However, if your application is
running on a single HP NonStop system, you will have to plan
for downtimes of 15 minutes or more at least every few years:
After all, even an HP NonStop system needs to be shut down
if you need an operating system upgrade, you’re switching to
completely new hardware, or you’re moving your data center.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 22 HP NonStop For Dummies, comForte Special Edition

The solution for avoiding any downtime in those cases comes

in the form of a so-called active-active system: You build a
cluster of at least two nodes in the following fashion:

✓ Both nodes always process transactions, normally shar-

ing the load.
✓ Both nodes carry a full local copy of the same database,
which are kept in sync. The synchronization is done using
replication software, typically available from third parties.
✓ If one of the two nodes fails, the other takes up process-
ing of all transactions.

Active-active isn’t the only way to achieve high levels of avail-

ability in case of planned downtimes for operating system
upgrades, hardware replacements, or the like. Hot standby,
where a ready-to-go hardware set is kept up-to-date just in
case, is often just as good or even better for some implemen-
tations. It is certainly easier to set up than true active-active.

An active-active configuration for an application is difficult to

build — but it can be and has been done on HP NonStop for
decades. Beyond uninterrupted uptime for up to decades (no
kidding!), it also gives you a perfect business continuity solu-
tion because the two nodes can be very far apart from each
other — for instance, one at the west coast and one at the
east coast of the U.S.

Key Benefits of HP

NonStop Systems
Here are some of the key benefits of HP NonStop.

Common standards, uncommon

advantages
The open standards supported by HP NonStop are key to the
ease of use and integration of the system. The support of stan-
dard technologies such as Java, C++, JSON/REST, SOAP, SASH,
and others — all running in a self-managing and self-healing
cluster in a box — combine to give enterprises that depend
on HP NonStop uncommon advantages.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 2: A Platform Overview 23
What uncommon advantages? The most distinguishing fea-
tures of HP NonStop are RAS (reliability, availability, and servi-
cability; see Chapter 1) together with its linear scalability (see
earlier in this chapter).

The exceptional serviceability of HP NonStop means that with

your application running and actively processing transactions,
you can replace just about any piece of hardware in the system.
This goes far beyond being able to replace a single disk in a
RAID array. Also, you can update several layers of software
(both at system and application level) without having to take
the application down.

In the financial world, it is totally normal for HP NonStop systems

to process several thousand transactions per second — for
hours and days and years on end, without missing a single
transaction. Need to process even more transactions? Thanks
to the linear scalability, you can simply add more hardware
capacity without needing to worry about hitting any system or
architectural limits.

Low TCO (total cost of ownership)

Comparing the smallest NonStop system with an industry
standard server cluster running Linux or Windows and a
commercial database initially often makes HP NonStop look
expensive.

However, add database clustering capability, extended fault

tolerance, and top-notch support to the standard server and
the picture quickly looks different.

Also factoring in the cost of managing the systems, the low TCO
of HP NonStop systems has been proven over and over again.
See www.comforte.com/ns4d/tco_study.pdf for a white
paper containing a detailed TCO study.

Excellent support
HP offers excellent customer support. By calling the HP global
support center and declaring a case as severe, you will get
a knowledgeable person talking to you in a few hours. This
person will not stop working on your issue until it is resolved
at least via a work-around.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 24 HP NonStop For Dummies, comForte Special Edition

The infrastructure for this has been built up over decades and
works like a charm to this day.

Cloud/hybrid computing
An HP NonStop system can be viewed as a cloud within a
single cabinet extendable to other cabinets and into commer-
cial clouds like Amazon, Microsoft, and Rackspace.

One good example of HP NonStop running in a cloud comput-

ing context is an implementation where HP NonStop serves as
the database backend for a web application provisioned in a
private or public cloud. This gives you the benefit of combin-
ing a cheap, scalable, and stateless front end (in the cloud)
with a reliable backend that is always there to keep state of
transactions.

Architectures like this are also called hybrid because multiple

pieces (also called tiers) of the application run on multiple,
different platforms. Mixing platforms like this allows each
platform to do what it does best and to also optimize for cost.

Hybrid architectures like this have been running, fully in pro-

duction, for various HP NonStop customers worldwide before
the term cloud computing was even coined.

HP NonStop on X86
Meg Whitman, the chair of Hewlett Packard, has announced
that the HP NonStop platform will be modified to run on the
X86 chipset without compromising any of the attributes that
HP NonStop stands for. When running on X86, NonStop sys-
tems will run on commodity HP server hardware, meaning
that the same hardware could, in theory, be running either
Windows, Linux, or HP NonStop.

Details for time frame, features, and price point aren’t avail-
able yet. But HP NonStop will continue to play an important
role in HP’s strategy for mission-critical computing.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 3

Developing Sof t ware for

HP NonStop
In This Chapter
▶ Using the tools you know to build HP NonStop applications
▶ Modernizing applications
▶ Migrating applications

N o matter whether you’re building new applications to

run on HP NonStop, maintaining or improving exist-
ing ones, or migrating applications to the platform, you’ll
find that the development life cycle isn’t much different from
what most developers are used to. Today’s languages and
debuggers can be used, cross-compilers can make testing and
deployment simple, and developers who never thought they
could build for what they consider mainframe-type or legacy
systems will do just fine.

Developing for HP NonStop

Doesn’t Mean Developing
on HP NonStop
Once upon a time when Tandem was young, Tandem’s engi-
neers invented a new programming language called TAL, and
for some time, TAL was the only language available for this
platform.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 26 HP NonStop For Dummies, comForte Special Edition

Those days are long gone. Although TAL is still available if

someone really needs it — typically for maintaining very old
applications — no new development for HP NonStop needs to
be done in TAL.

Today, all standard development languages, such as C, C++,

COBOL, Java, Python, and Perl are available on and for the
platform. We look at some in more detail in this chapter.

Programming languages and

development environments
When developing for Windows computers, developers work on
Windows computers. The same is true of Macintosh and Linux.
But how about Android or iOS? No one develops on a phone,
right? They all use a common IDE like Visual Studio or Eclipse,
and a cross-compiler that builds, links, and deploys to the phone.

The same is true of HP NonStop. Compilers and linkers are

available for TAL, C, C++, and COBOL that compile, link, and
generate the objects needed to run the application on the big
iron. All of these have support for embedded and dynamic
SQL for SQL/MP and SQL/MX databases. The developer’s
IDE of choice (and often libraries of choice) can be used to
develop, cross-compile, and debug applications.

In a nutshell, there are two options to create an executable for

HP NonStop:

✓ Invoke the compiler right on HP NonStop. This has been

the only option for decades and it stays a viable option.
✓ Use cross compilers on a PC. For a long time, that meant
using a plug-in called ETK from HP for the Microsoft IDE.
HP is currently transitioning its cross-compilers to run
under the Eclipse IDE rather than under the Microsoft IDE.

Using Eclipse to develop

for HP NonStop
The front-runner of Enterprise IDEs is Eclipse (see Figure 3-1 for
a screenshot of a C++ project for HP NonStop being worked on
in Eclipse), and the popular environment is about to replace

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 3: Developing Software for HP NonStop 27
the Microsoft IDE as IDE of choice for development for HP
NonStop. HP has a plug-in for Eclipse called NSDEE that will help
the newer generation of developers feel right at home develop-
ing applications for HP NonStop.

Figure 3-1: H
 P NonStop development environment for Eclipse.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 28 HP NonStop For Dummies, comForte Special Edition

Enter Java
The Java programming language has been supported on the
HP NonStop platform for quite some time. Experienced Java
developers might question the production speed of Java in a
­fault-tolerant and message-based environment like HP NonStop.
Java has been clocked as fast on HP NonStop as it is on Unix,
but your individual mileage will vary depending on many
­factors, such as how well your Java application follows design
principles.

HP offers the SASH stack — consisting of Spring, Axis2, MyFaces,

and Hibernate — for the platform. This offering is no different
than what is found on any Unix box. If you have an applica-
tion that doesn’t require too many concurrent transactions,
spends most of its time in complicated SQL queries, and needs
sophisticated user interaction, Java using the SASH stack on HP
NonStop may be absolutely the right choice for you.

Also, the product called HP NonStop Application Server for

Java, or NSASJ, consists of a special version of JBoss, internally
built with Pathway. There are JARs available that make integra-
tion of NSASJ into your application simple and reliable. This
dramatically increases manageability of projects written in Java,
and also shows HP’s commitment to Java as a go-forward devel-
opment language for HP NonStop. While a standard JBoss imple-
mentation on Unix/Linux uses a low-level interface called sockets
to communicate between processes, the HP engineers replaced
that layer with the ultrareliable and fast Pathway interfaces.

By doing this, they automatically also got TMF integration. This

is invisible to the developer and makes your standard JBoss
application take full advantage of NonStop’s scalability and data
integrity features. HP also replaced the need for shared memory
within JBoss by a fault-tolerant and distributed context-keeping
process, so that JBoss applications transparently can be distrib-
uted over several shared-nothing CPUs.

Debugging is in the mix

All these development tools are great if you write bug-free
code. You probably know some people who don’t write bug-
free code, however, and therefore you can debug your code
running on HP NonStop in a variety of ways:

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 3: Developing Software for HP NonStop 29
✓ If you compile on the HP NonStop, you can debug from a
shell prompt on the system.
✓ If you use NSDEE from Eclipse IDE, you can use the build-
in debugger from Eclipse (see Figure 3-1).

Writing Good Software

for HP NonStop
Although the HP NonStop OS does provide its core benefits
out of the box, some design principles have to be taken into
account for an actual application to fully leverage this.

Assembling a team of developers

Getting a team of developers together is easier than it must
seem at first, because of the available tools for the platform
and the fact that all modern development languages are sup-
ported. This is as it should be — there is a straightforward
pattern to building a development team for a project of
this scope.

What it comes down to is using the best developers in your

chosen development environment and language(s). Then,
find one or two HP NonStop specialists to engage at the archi-
tecture and configuration tier of the effort and make sure the
code will perform at the level you expect.

The biggest factor to success on an HP NonStop project is no

different than on any other platform: Let senior people who
truly understand the requirements as well as the architectural
options on HP NonStop run the show. System architecture and
best practice can be described by a few specialists, even on a
large project.

Writing efficient code

Writing the most efficient code possible improves the perfor-
mance of HP NonStop applications just as efficient code does
in any other environment. Efficiency comes down to through-
put, response time, and resource consumption. Balancing

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 30 HP NonStop For Dummies, comForte Special Edition

these requirements appropriately for the problem at hand

leads to efficient systems that perform well.

Writing efficient code on HP NonStop comes down to under-

standing and making use of the platform fundamentals. It will
not require you to use platform-specific features such as non-
stop programming or no-waited programming (see next sec-
tions). We discuss this in more depth in Chapter 5.

Nonstop programming
If you deem it necessary, there is the possibility to write your
very own code implementing a process as a process pair.

However, this is never required or even recommended for an

application program — using development frameworks such
as Pathway or NSASJ/JBoss will take care of fault tolerance for
you without having to write any special code.

No-waited programming
One important — and somewhat unique — technology available
on HP NonStop is called no-waited programming. No-waited pro-
gramming appears on the surface to be a lot like so-called asyn-
chronous programming on other platforms but some significant
differences exist. Fortunately, this is only needed in rather rare
circumstances of low-level systems programming (see Chapter 5
for more details).

Modernizing Existing
Applications on 
HP NonStop
Many applications running on HP NonStop have been writ-
ten years if not decades ago, many of them in TAL or COBOL.
Although they’re still providing the functionality they were
written for, they often lack a modern GUI or contemporary
security features, good integration with the Enterprise archi-
tecture, or simply the capability for new development in
modern languages.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 3: Developing Software for HP NonStop 31
Fortunately, there are many options for modernizing legacy
applications, which often provide a better ROI and less risk
than replacing applications that contain hundreds if not
thousands of years of stable code (for more on this, see
Chapter 5).

Porting Existing Applications

to HP NonStop
New development on HP NonStop has become more modern
and approachable, but how about migrating existing applica-
tions to the platform? Applications that require new levels of
reliability often need to change platforms to achieve total fault
tolerance — wouldn’t it be nice if the platform switch could
be done without a lot of code changes?

The best way to look at this is through example. In 2011, HP

performed a migration of a real-time fraud detection system. It
was written about in detail within a case study you can find at
www.comforte.com/ns4d/migration.pdf. Speed and fault
tolerance were incredibly important because the service was
used to check transactions as they were processed for fraud.

The original application was built on Oracle, developed in

Java using the Spring and Hibernate frameworks. Because HP
NonStop embraces these technologies in modern versions of
the platform, migration was smooth. HP JToolkit and Pathway
were used to improve overall reliability and manageability.

Database migration
Oracle migration to SQL/MX posed few challenges because
the original architect of the application kept things simple —
proving again that the old keep it simple rule holds under
many different circumstances. Oracle data types map well to
SQL/MX data types, as shown in Table 3-1.

JDBC was used in the reference application for data access

to the Hibernate schema. For the OSS application, there is a
Hibernate dialect jar, called hibernate3sqlmx, which provides
nearly all the same features.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 32 HP NonStop For Dummies, comForte Special Edition

Table 3-1 Mapping Some Oracle Data Types

to SQL/MX
Oracle SQL/MX
Char Char
VarChar2 VarChar
Date TimeStamp(0)
Number (<5 digits) Smallint
Number (<10 digits) Integer
Number (<15 digits) Largeint
Integer Numeric(38)
Number(10,3) Numeric(10,3)
BLOB VarChar

Application migration
The reference application was built using the Model-View-
Controller (MVC) pattern, greatly exposing separation
between business logic and data. This design paradigm pro-
vided an ease of migration to the application layer similar to
what was found in the data layer.

The port replaced standard communication methods (sockets)

with HP NonStop-specific methods ($RECEIVE-based messag-
ing and Pathway). Due to the modular architecture of the Java
Spring framework, changes were required only at a few central
places. The migrated code takes full advantage of the transac-
tional architecture of HP NonStop.

Further considerations
In the past few years, several applications have been suc-
cessfully ported onto the HP NonStop platform. HP as well as
the independent software houses (ISVs) provide a wealth on
expertise on how to do this smartly.

Breaking an application into multiple tiers and running these

on multiple platforms can also eliminate or simplify the
requirement to port the whole application: The parts of the
application that remain running on their initial platform will
need very little if any porting work.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 4

HP NonStop as Par t of
Enterprise IT
In This Chapter
▶ Securing HP NonStop systems properly
▶ Integrating with other Enterprise IT systems
▶ Examining the modernization of the platform and/or applications

I ntegrating HP NonStop hardware and software with the

rest of your enterprise IT infrastructure requires ­careful
planning, but is easier than ever before. HP NonStop has
options for REST and SOAP service integration, for MQ mes-
saging, and for other intersystem communication options.

Security is first and foremost in the system architecture, pro-

tecting the interactions with other entities. HP NonStop is a
first-class citizen of the enterprise environment.

Securing HP NonStop Systems

In general, securing HP NonStop is no different than securing
any other server at the core. When thinking about security for
NonStop, you should remember a few things, though:

✓ NonStop applications are mission critical — a failure

because of a security flaw could be a disaster.
✓ NonStop systems very often carry extremely sensitive
data like credit card transactions or medical information.
Cost must not be an issue when protecting this data.
✓ From a low-level perspective, NonStop systems are very
different from the rest of IT. Security considerations must

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 34 HP NonStop For Dummies, comForte Special Edition

consider this. For instance, a virus scanner makes abso-

lutely no sense on NonStop. However, database content
encryption can make a lot of sense.

To make NonStop secure, you must consider at least these

three facts. But although this list is short, not many people
really understand how to do this. Hire experts and let them do
their work. Your sleep will be better.

Security has come a long way

In the last ten years or so, the built-in security of the HP
NonStop platform has come a very long way by adding new
capabilities, such as encryption of data in transit, strong
authentication, or integrating them as the source for central
event logs (SIEM systems). Depending on the specific imple-
mentation and configuration of an individual system, the
common security attack vectors still exist, of course, includ-
ing but not limited to:

✓ Using application security flaws — for instance, SQL

Injection
✓ Network sniffing
✓ Attacks against media
✓ Breaking access control

General platform hardening

HP NonStop has a built-in security framework called Safe­
guard. Safeguard is a powerful component baked deeply into
the NonStop operating system, which handles a lot of the
on-platform security by providing user management and the
­capability to configure Access Control Lists (ACLs) for files
and processes.

Then there is the list of details that aren’t handled by

Safeguard. They really aren’t that different than the things
you have to do to secure any other server:

✓ Replace the username/password authentication with SSH

certificates or federate with Active Directory or LDAP.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 4: HP NonStop as Part of Enterprise IT 35
✓ Let administrative users employ two-factor authentication.
✓ Encrypt data in transit with SSL or SSH.
✓ Encrypt data at rest at the application level or by using
products that encrypt the underlying database without
requiring application changes.
✓ Maintain file integrity and check system and application
object files.
✓ Given the low number of HP NonStop installations com-
pared to Windows, Unix, and Apple, there has not been a
single instance of malware reported on the platform. For
that reason, there is no malware scanning tool available.
✓ Apply reasonable security standards, especially PCI and
HIPAA. Be sure to apply the spirit of the standard, not
just the letters.

Specific hardening of the

HP NonStop platform
The uniqueness of the platform results in having to secure
components (such as Pathway) that aren’t described in your
typical server hardening document. For some resources spe-
cific to the platform, see Chapter 5.

IAM: Identity and access

management
Large central servers are often in tightly controlled environ-
ments, with few users and even fewer administrators. They
also last a long time, which means that as users come and go,
user accounts often don’t. IAM is something that needs to be
considered as a process when dealing with HP NonStop.

Identity federation is the best overall solution for IAM on

HP NonStop. Leveraging Windows authentication to access
the HP NonStop system provides central management and
can potentially also provide true single-sign-on. It is also
strongly recommended to use two-factor authentication to
get access to HP NonStop rather than relying on username
and password alone.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 36 HP NonStop For Dummies, comForte Special Edition

Sharing usernames? Don’t even think about it.

Another impact of the set it and forget it factor of large serv-

ers is that log files get lost in the ether. Safeguard, HP’s log-
ging system for the Guardian platform, does an excellent job
of logging user events. It’s highly recommended that those
logs make it into a Security Information Event Management
(SIEM) system like HP ArcSight. Other exception logs, such as
application-specific log files and the EMS messages, should
also be sent to a SIEM system.

Keys and certificates

Management of keys and certificates for SSL, SSH, and data-
at-rest encryption is no different on HP NonStop than on
any other system. As always, maintenance and protection of
secret assets is the key, if you’ll excuse the pun.

Existing corporate Public Key Infrastructure (PKI) systems

should be extended to support HP NonStop applications, just
as they should with any other encryption or messaging effort.
Implementation of encryption is a one-off programming or instal-
lation task, but key management is an ongoing maintenance
task. Make use of your existing talents and processes there.

Finally, the SSL and SSH protocols both only protect against
sniffing attacks if at least server authentication is configured
correctly on both the server (typically HP NonStop) and the
client (typically a PC connecting to HP NonStop).

Integrating HP NonStop
with Other Systems
For decades, HP NonStop systems were the beginning and
the end of the chain of applications that resided there. The
program ran on the NonStop and you accessed it with a dumb
terminal — it did its job and you did yours. That’s it.

Doesn’t work that way today, though, does it? Applications

and work flows span multiple systems, data has to be acces-
sible from anywhere, and there are four different form factors
for every user interface. Integration isn’t just important, it’s a
requirement.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 4: HP NonStop as Part of Enterprise IT 37
Client-server and multitier design
NonStop lends itself very nicely to multitier applications.
Although traditional HP NonStop applications were built
entirely on the one platform, today’s applications are expected
to have the user interface rendered by various other systems,
or even handled by the client.

HP NonStop’s core architecture is perfect for this. Unlike some

monolithic applications where data, business logic, and user
interface code are all batched together, HP NonStop applica-
tions have separated these tasks since the beginning.

The trend toward apps running on mobile devices is, from

a technical viewpoint, just an extension of client/server. It
can be accommodated by distributed computing where the
NonStop typically will play the role of the transactional data-
base backend.

Raw messaging
One of the reasons that business logic and data are sepa-
rated in HP NonStop applications is the strong history of
messaging on the platform. Raw binary encoded messaging
has been implemented since the beginning, either as part of
Pathway application design or using both custom developed
endpoints and commercial products like the HP Remote
Server Call (RSC) product.

In today’s environment, binary encoded messaging isn’t

enough. New messaging protocols have made integration
much easier.

Simple Object Access

Protocol (SOAP)
SOAP is a rich XML-based message passing protocol using an
envelope model for message metadata. SOAP supports many
enterprise-required features out of the box, such as reliable
messaging, privacy, and transaction support.

HP NonStop systems do come with the HP NonStop SOAP

product; also, many Java toolkits have a rich set of SOAP

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 38 HP NonStop For Dummies, comForte Special Edition

capabilities. And you can find a number of commercial prod-

ucts out there that provide additional features compared to
HP NonStop SOAP.

Representational State
Transfer (REST)
REST is the new buzzword and uses the well-established
HTTP protocol for message passing. Usually using JavaScript
Object Notation (JSON) for data transfer, REST takes advan-
tage of HTTP verbs for CRUD operations:

✓ PUT: Create
✓ GET: Read
✓ POST: Update
✓ DELETE: Delete

REST is usually used for quicker, lighter applications than

SOAP, because it doesn’t have a thick infrastructure for reli-
able messaging or security — just what is provided by HTTP.
For straightforward CRUD-style interaction with a system’s
backend, however, REST is an excellent messaging protocol.

REST has become the de facto standard for mobile applica-

tions. Support within HP NonStop is significant because of the
strong presence of mobile and fault-tolerant systems in tele-
communications and mobile payment systems.

REST-based frameworks are available for NonStop from sev-

eral third parties, and they give application developers an
awesome, reliable way to integrate HP NonStop servers into
modern application designs.

Operations Management
Watching what a system does is an important step to pre-
vent the unexpected and undesired. EMS is the main system
event message hub for the HP NonStop (for more on this, see
Chapter 2). However, several commercial products are avail-
able that allow you to monitor an HP NonStop system together
with other systems, under a single operative umbrella.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 4: HP NonStop as Part of Enterprise IT 39
To name a few:

✓ Nagios is an open source tool that has been successfully

adapted to monitoring HP NonStop systems.
✓ HP itself offers several tools; visit the HP NonStop web-
site (see Chapter 5 for this) and search for “HP NonStop
manageability” products.
✓ The Tivoli suite from IBM happily interoperates with HP
NonStop.
✓ There are several third-party (“ISV”, see Chapter 5) prod-
ucts that provide powerful capabilities.

You can find a number of products that put a GUI on top of the
command-line or block-mode screen-based tools HP NonStop
offers out of the box. Refer to Figure 2-1 to see a screen shot
of one of these.

File Transfer
Moving files around is arguably very important in enterprise
integration. File Transfer Protocol (FTP) is the venerable pro-
tocol developed for Internet connected computers, and is, of
course, supported on HP NonStop.

These days, FTP has been superseded by SFTP or SFTP/SSL,

which are the secure file transfer protocols that are gradually
replacing FTP. Fortunately, both are supported on HP NonStop.

If you need advanced features such as guaranteed file deliv-

ery, there are several commercial products available that
implement what is then called managed file transfer.

Messaging
In Chapter 2, we discuss the fact that the HP NonStop OS is
message-based at its core. In this section, we discuss messag-
ing in a completely different context.

Asynchronous messaging means that applications send mes-

sages without waiting for an immediate reply. These messages
are typically stored in a database, but only for a short period

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 40 HP NonStop For Dummies, comForte Special Edition

of time until they can be delivered to the final destination (or

to the next station if they have to travel a long way). Email
systems work that way, but in the NonStop world, messages
are typically sent from application to application — where
applications can run on different platforms. These kinds of
messaging applications are used for quite different purposes
on NonStop:

✓ Manufacturing devices send status or work progress

messages to central servers.
✓ Banking applications send messages to other banks.
✓ Credit card processing applications send messages to
central billing systems.
✓ Many other examples are possible.

The common properties are that the messaging must be reli-

able, messages must be sent exactly once, the sequence of
messages is typically important, and the application doesn’t
track the message. The underlying infrastructure where the
application put the message must be so reliable that there is
no need to care after the message has been put.

Messaging is very important in modern applications. For

Java-based applications, there is the JMS standard defining a
common, standardized interface how messages can be sent
and received. As JMS is only an interface standard, there
may be many different implementations, even on the same
platform.

For NonStop systems, many industry-strength messaging

applications are available right now:

✓ There is a JMS implementation based on SQL/MX avail-

able from HP (mainly for local messaging).
✓ Another JMS implementation based on Apache Active
MQ is also available from HP (product name NSMQ).
✓ Websphere MQ, IBM’s very popular messaging product,
is available for NonStop from IBM.
✓ Finally, other third parties might also provide messaging
solutions for NonStop.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 5

Ten Resources for More

Information
In This Chapter
▶ Ten resources for HP NonStop information

T here is more to the HP NonStop platform than is possible

to fit into a little bitty book like this. The tome in your
hands is a good primer, but you probably have more ques-
tions. In this chapter, you find links with a lot of additional
information to help you find out even more. From developing
to auditing to securing, these references are your next step to
expanding your knowledge of HP NonStop.

NonStop Terminology and Entities:

A Quick Reference
Throughout this book, we include terms that may be new to
you, such as TACL, TAL, Pathway, Safeguard, SCF, and EMS.
Although we could not squeeze a reference section into this
book, you can find some extra information on these and other
terms at www.comforte.com/ns4d/quickref.

TACL Fact Sheet

Do you know your way around the Windows CMD prompt
and/or the Linux/Unix Shell and want to quickly find a com-
mand? Fear no more; the table at www.comforte.com/ns4d/
taclcheatsheet will help you.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 42 HP NonStop For Dummies, comForte Special Edition

Open Source Tools Available

on HP NonStop
We mention open source tools in several locations in this
book, but we do not provide either potential sources or a list
of available tools. If you are interested in these details, please
see www.comforte.com/ns4d/opensource.

Tips for Developing
for HP NonStop
Developing for HP NonStop uses the same languages, IDEs,
and skills as contemporary applications (this is also discussed
in Chapter 3). Developers will find that the life cycle that
they’re familiar with still applies in nearly every case. That
said, some tips will make the platform feel more functional
and reasonable. Find these tips at www.comforte.com/
ns4d/devtips.

No-waited Programming
We briefly mention no-waited programming in Chapter 3. In a
nutshell, this is required only for system-level programming
where your application wants to handle multiple events in
parallel. If you want to know all the gory details, find them at
www.comforte.com/ns4d/nowaited.

Modernizing Legacy
Applications
Although the HP NonStop platform is very modern, some
applications would benefit from a little refresher. An abun-
dance of options are discussed in more depth at www.
comforte.com/ns4d/modernize.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 Chapter 5: Ten Resources for More Information 43

Securing NonStop Systems

Many HP NonStop users run the platform to process credit
and debit card data — either through connected point of sales
(POS) systems or through ATMs. In both cases, the HP NonStop
is at the core of the architecture, making it a prime target for
cybercriminals. Because of this association with the financial
industry, many HP NonStop customers fall under the Payment
Card Industry Data Security Standard (PCI-DSS).

Although your security architecture should ideally not be

driven by a single regulatory framework, the PCI-DSS is widely
regarded as a well-worded guideline to make sure you have an
appropriate security posture.

For more resources on the topic of PCI on NonStop as well as

on securing NonStop systems in general, please go to www.
comforte.com/ns4d/security4d.

The NonStop ISV Ecosystem

Old-timers in the NonStop (Tandem) world know what an
ISV is. ISV stands for independent software vendor and simply
means a software house that isn’t part of HP and which devel-
ops and markets software specific for HP NonStop systems.
comForte, the company behind this book, is an ISV.

There are plenty of these around, and many of them are listed
on the HP website (see next section) under partners. Also,
you will see them as sponsors of some of the User community
meetings (see section “The User Community”).

ISV applications and system software play an important role

to make the HP NonStop platform attractive — without soft-
ware and tools, the best hardware has only very limited use.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 44 HP NonStop For Dummies, comForte Special Edition

The HP Website and HP

NonStop Computing
Hewlett Packard is a large organization with more than
300,000 employees working on a huge set of products and
services. Designing a web page to inform your customers and
to offer support for both humble printers for home usage
and high-end systems like HP NonStop is a challenge.

However, you can find a lot of great resources about HP NonStop

on the HP web page. The URL www.hp.com/go/nonstop is a
helpful shortcut to get to these pages, which we recommend
highly.

Also, the URL www.hp.com/go/nonstop-docs will take you

right to the extensive documentation for HP NonStop systems.

The User Community

Most people dealing with HP NonStop systems are passionate
about the platform. Ever since the start, there was always a
strong community around the technology, supported by HP
(Tandem) and the ISVs, but really driven by people who know
that technology alone doesn’t make projects succeed. Sharing
knowledge about technology is a great help, however, and
that’s what this community is all about.

Today the community is strong as ever with online resources,

a printed bimonthly magazine, and regular user group meet-
ings from small local meetings to large gatherings. See www.
comforte.com/ns4d/community for more.

These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
These materials are © 2014 John Wiley & Sons, Inc. Any dissemination, distribution, or unauthorized use is strictly prohibited.
 WILEY END USER LICENSE
AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s
­ebook EULA.