Hitachi Virtual Storage Platform Family

Advanced Storage Capabilities for All Organizations

Andre Lahrmann
23. November 2017
Hitachi Vantara

Vorweg:

Aus Hitachi Data Systems wird

Hitachi Vantara
The efficiency
you deserve
Agenda

Overview Virtual Storage Platform

Hardware Architecture

Flash Strategy and Portfolio

Software Architecture and specific Features

The efficiency
you deserve
Agenda

Overview Virtual Storage Platform

Hardware Architecture

Flash Strategy and Portfolio

Software Architecture and specific Features

The efficiency
you deserve
Hitachi Virtual Storage Platform (VSP)

Leading performance to accelerate business

All-flash and hybrid solutions with up to 4.8 million IOPS

Built-in efficiencies for improved total cost of ownership (TCO)

No penalty flash compression and high-speed deduplication

Powerful automation to simplify IT

Provision and deliver resources in a fraction of the time

Proven resiliency for peace of mind

Eliminate disruption and guarantee data access

Best-in-industry virtualization for IT agility

Easily update resources without changing process
Choose the Right System

VSP G The most powerful enterprise array available

Unmatched performance, automation and resiliency
Series Maximum flexibility and choice in configuration

VSP F The most resilient all-flash array on the market

Intelligence and performance to power tomorrow’s IT
Series The easiest way to add all-flash to a Hitachi environment
 Portfolio Positioning – Virtual Storage Platforms

COMMON SOFTWARE AND MANAGEMENT

Hitachi Storage Hybrid Arrays
Virtualization
Operating System
(SVOS7)
Feature Set
Across Family
G1500
G1000 Fully Supported
in Hitachi
G800 Command Suite

G600
AFA-All Flash Arrays
G400
F400 F600 F800 F1500
G200

COMMON OPERATING SYSTEM

 VSP G Series Overview

Controller 2U Controller 4U Controller 4U Controller 4U Controller 10U

FE Ports 16 FE Ports 64 FE Ports 64 FE Ports 80 FE Ports 192

Drives 264 Drives 480 Drives 720 Drives 1440 Drives 2304

Capacity 2.5PB Capacity 5PB Capacity 7PB Capacity 14PB Capacity 8PB

Cache 64GB Cache 128GB Cache 256GB Cache 512GB Cache 2TB

Volumes 2,048 Volumes 4,096 Volumes 4,096 Volumes 16,384 Volumes 65,280

IOPS 270K IOPS 600K IOPS 800K IOPS 1.4M IOPS 4.8M
VSP F Series Overview

Controller 4U Controller 4U Controller 4U Controller** 10U

FE Ports 32 FE Ports 32 FE Ports 48 FE Ports 128
Flash Modules 192 Flash Modules 288 Flash Modules 576 Flash Modules 576
Capacity (5:1) 13PB Capacity (5:1) 20PB Capacity (5:1) 40PB Capacity (5:1) 40PB
Cache 128GB Cache 256GB Cache 512GB Cache 2TB
Volumes 4,096 Volumes 4,096 Volumes 16,384 Volumes 65,280
IOPS 600K IOPS 800K IOPS 1.4M IOPS 4.8M
Agenda

Overview Virtual Storage Platform

Hardware Architecture

Flash Strategy and Portfolio

Software Architecture and specific Features

The efficiency
you deserve
VSP G200, VSP G400, VSP G600 and VSP G800

VSP G400, VSP G600, VSP G800

VSP G200
 Hardware Overview
2U x 24
SSD (SFF)
SAS (SFF)

2U x 12

Fibre Channel:
16G, 8G NLSAS (LFF)

iSCSI (SFP):
4U: 400/600/800 SSD (SFF)
10G 4U x 60
2U: G200
iSCSI (10GBase-T): SAS (SFF)
10G

2U x 12

FMD (LFF)
Available HDDs

Type HDD
SSD 480GB
960GB
1,9TB
3,8TB
SAS 600GB
1,2TB
1,8TB
NL-SAS 4TB
6TB
10TB
FMD 3,5TB
7TB
14TB
VSP Gxxx File Blade Specifications

10G 10G IOPS/Throughput VSP G400 VSP G600 VSP G800

Cluster NFS & 3 GB/s 4 GB/s 10 GB/s
NFS & CIFS Read Throughput
Ports CIFS
Ports NFS & CIFS Write Throughput 1.5 GB/s 2 GB/s 5 GB/s
NFS & CIFS Mixed (70% read & 30%
2 GB/s 3 GB/s 6 GB/s
write) Throughput
NFS Spec2008 in IOPS 140,000 200,000 270,000
CIFS Spec2008 in IOPS 140,000 200,000 270,000
Scalability
Max and Min blades per cluster! 2
Max storage per entity 5 PB
Max file system size 256 TB
Max # of FS per Cluster 125
Concurrent Open Files 96,000 512,000 1,024,000
Concurrent Connections 36,000 64,000 64,000
Port Speed
Network 10 GbE 4 x Active 6 x Active
12 x 8GB DDR3 DIMM Cluster 10 GbE 2 x Active 4 x Active
96 GB memory
exclusively for NFS and
CIFS operations
File Blade VSP G400, VSP G600 and VSP G800

VSP G400, VSP G600, VSP G800

Agenda

Overview Virtual Storage Platform

Hardware Architecture

Flash Strategy and Portfolio

Software Architecture and specific Features

The efficiency
you deserve
CPU Performance Development

MIPS CPU growth

1.000.000,00

100.000,00

10.000,00

1.000,00 1980 until now

Appr. 300.000x-times
100,00
increase

10,00

1,00

0,10 Source: Wikipedia

1970 1975 1980 1985 1990 1995 2000 2005 2010 2015
HDD Capacity Growth vs. Performance

HD THROUGHPUT MB/S
300
270

250

200
200

150

100
64

50 36

2,5 4 77,6 9,5

0
1985 1990 1995 2000 2005 2010 2015 2020

Max. HDD speed

15k UPM since 15 years constantly

Source: Wikipedia, University of California, Berkeley

What Will Drive the Flash Trend?

CONTINUING
FLASH PRICE
REDUCTIONS
 Enterprise Flash Deployment Growth

Flash has Already Replaced HDDs

2,9%
for Block-based Storage

Within the Next Year 3,9%

Within the Next 1-2 years 12,4%

Within your own organization,

when do you believe all-flash Within the Next 2-4 years 32,7%

approaches will completely

replace traditional hard disk- Beyond 4 Years 35,8%

drive based approaches for

block-based (i.e. SAN-based) Never 12,4%

storage workloads? n = 590 Percent of Sample

Source: 451 Research, Voice of the Enterprise: Storage, Q1 2016

Hitachi’s FLASH INNOVATION

FLASH ONLY ARRAYS

BLOCK / PAGE MAPPING
PERFORMANCE MANAGER
WEAR LEVELLING
ZERO BLOCK COMPRESSION
DATA REFRESH
ENDURANCE MANAGER

KEEP THE ENTERPRISE FUNCTIONALITY THAT WORKS

OFFLOAD FLASH MANAGEMENT FUNCTIONS TO
HARDWARE
Hitachi Innovation: Flash Module Drives

High Flash High

Battery
Density Controller Density
ASIC Backup
Flash Flash
New 14TB and 7TB FMD HD

Built on 15nm NAND (128GB) flash

Industry leading 14TB capacity option

‒ Targeting large capacity systems
‒ Available on all VSP G/F

With 2:1 compression, up to 28TB

effective capacity each

A single 7+1 represents ~196TB effective

capacity
Hitachi Flash Module (FMD) Advantage

Accelerates more workloads with fewer resources

‒ Up to 3x the read, 5x the write performance of SSDs
‒ Delivers 4x the system throughput (24Gb/sec) Parallel Block/Page
of competitors Connections Mapping

Increase capacity with no performance penalty Contiguous

Wear
Block Write
Leveling
‒ Uses inline compression 10x faster than competitors Avoidance

‒ Provides 4x more effective capacity for a lower TCO Proactive Extended

Monitoring ECC
Including
Write
‒ 20% Spare Flash, Self Healing, Write Depletion Monitoring Data
Capacity
Compression
Monitoring
Agenda

Overview Virtual Storage Platform

Hardware Architecture

Flash Strategy and Portfolio

Software Architecture and specific Features

The efficiency
you deserve
SVOS – Storage Virtualization OS

Virtual
‒ Virtual LDEVs for hardware abstraction and
implementation of an unified virtual storage layer
‒ Volume Migration
‒ Virtual storage machines and external storage
virtualization

Available
‒ 100% Data Availability Warranty
‒ Non-disruptive Microcode-Update
‒ Remote-Replication
‒ Global-active device and nondisruptive data migration
 Hitachi Storage Virtualization Operating System
Virtualizes Externally Attached Storage

Fibre Channel
or iSCSI
Target Port
ELUN
Host View
Devices

Servers LUN

SVOS Controller Intelligence

VDEV VDEV VDEV
Devices
Internal

LDEV
LDEV

Mapped Mapped
1:1 External Port N:1
Target Port
Internal
Disks

Target Port
HDD

External
Storage
LUN LUN LUN
RAID group
LUN LUN LUN
 VSP Gx00 MicroCode Update

Write Cache is NOT disabled

Host
during Micro-Update because
Controller-based
the MP could be updated one by one or in groups.
Lun 0 Storage Virtualization
with 2 DC GAD
Two fully redundant Two fully redundant
Storage Controller Storage Controller
with integrated Storage Lun 0 with integrated Storage
Virtualization Softw.
Lun 0
Transparent Failover Virtualization Softw.

Synchronous Mirroring

Quorum
Internal Storage

External Storage

Microcode Maintenance Updates without loss of redundance or Interuption to Mirroring or Write Caching
Hitachi Global Storage Virtualization
Clustered Active-Active Systems

Global Storage Virtualization

Global Active Device
Virtual LDEVs:
00:01
00:02
Servers with Apps Servers with Apps
Requiring Virtual Storage Virtual Storage Requiring
High Availability Identity 123456 Identity 123456 High Availability

Resource Resource
Group 1 Group 2
Write to
Multiple Copies LDEVs: LDEVs: Read Locally
10:00 20:00
10:01 20:01
Simultaneous 10:02 20:02 Simultaneous
from Multiple from Multiple
Applications Applications
GAD Enhancements

Active-Active for every LDEV

Transparent Failover and Failback for Applications

Configurable with Shortest Way for Hosts

Needs a Quorum (virtual or physical) per iSCSI or FC

SVOS – Storage Virtualization OS

Secure
‒ Data at rest encryption
‒ Volume Shredder

Efficient
‒ Thin Provisioning and Disk Pooling
‒ Hitachi Dynamic Tiering
‒ Deduplication and Compression (Flash)

Fast
‒ Flash performance acceleration and optimization
‒ Quality of service (QoS) for consistent performance
Active Flash - Performance & Efficiency Evolution

HDP HDT Active flash

-Thin provisioning for efficiency -Automatically moves the most Seconds to sub-seconds
-Wide striping for performance active data to the fastest tier for
-Fixed workload assignment max performance; least active to
cycle time!
-HDT with LOWER RESPONSE TIME
lower cost tiers capabilities
-30 min to 24 hour cycle time -Inherits HDT & HDP advantages
DP Vol 1 DP Vol 2 DP Vol 3 -Inherits HDP advantages
HDT Pool 2 DP Vol 2
HDT Pool 1 DP Vol 1

Flash POOL Active flash: Rapidly-active Data

Tier 0 Flash Tier 0 Flash
SAS POOL Tier 1 SAS Tier 1 SAS
NL-SAS POOL Tier 2 NL-SAS Tier 2
NL-SAS
Deduplication and Compression

Intelligently adjusts mode to minimize latency 1

Incoming I/O
written to cache
‒ SVOS detects if data pattern already exists
‒ Compresses existing data inline
‒ Final de-duplication is handled post process Data analyzed for
duplicates
CACHE Updated data is
Optimizes data handling for performance compressed in-
line
‒ Application I/O prioritized for customer experience
‒ Smart selection prioritizes “stable” data first New data is
written to storage

Manages services used based on configuration

‒ If FMDs detected, FMD compression used PERSISTENT STORAGE
‒ If encryption required, SVOS compression used
SVOS intelligently
determines when to use
FMD compression.
SVOS 7: Enhance Customer Experiences

Quality of service (QoS) for consistent performance

‒ Prevents workloads from monopolizing bandwidth
‒ Prevents monopolization of data reduction cycles QoS: Predictable
‒ Integrates with VMware Virtual Volumes (VVols) for user experience
end-to-end QoS
SVOS: Flash- SVOS
VVol
Flash-aware I/O stack for long term, low latency aware low latency

‒ Prioritizes application I/O over background tasks

‒ Rebalances data placement to prevent hot spots

FMD

FMD

FMD

FMD
SVOS / FMD:
‒ Offloads tasks to FMDs for scalable low latency Managed flash
ongoing high
performance
Be Ready for Your Next Deployment Need

Operating Hypervisor
Cluster Support Certification
System Support Support
Questions and Discussion