007 Backup and Disaster Recovery
007 Backup and Disaster Recovery
Security Level:
Objectives
• Describe the technologies related to the storage backup solution of the data
center.
1 Huawei Confidential
Contents
1 Introduction to Business
Continuity
2 Introduction to Backup Solution
2 Huawei Confidential
Contents
1 Introduction to Business
Continuity
• Why Do We Need Business Continuity?
• Key Concept
3 Huawei Confidential
Why Do We Need Business Continuity?
Losses per hour due to system downtime
Sales and
648 19 17,000
marketing
280
200 Accounting 21 19,000
110 160
Unit: US$10,000
9 63
Engineering 42 98,000
Media Healthcare Retail Manufacturin Carrier Energy Finance
g
Source: Network Computing, the Meta Group and Contingency Planning Research
Source: Ontrack 2010
4 Huawei Confidential
Why Do We Need Business Continuity?
Disasters that undermine the business continuity of IT systems
Device failures
High
Data damages
Statistics show that:
Application interruptions Partial failures (97%)
Disk failures
cause interruption of
Yearly frequency
Component failures
Network failures enterprises' external
Regional incidents services.
Cooling system issues Global failures (3%)
cause data loss and
Power failures Terrorist/Domestic
DC failures Building even enterprise
turmoil
fire bankruptcy.
Natural disasters
Low
A disaster recovery (DR) system addresses the challenges of disasters and faults for
service continuity
5 Huawei Confidential
Key Concepts — BC and IA
Business continuity (BC) is an integrated and enterprise-wide process that includes all
activities (internal and external to IT) that a business must perform to mitigate the
impact of planned and unplanned outages.
The goal of a BC solution is to ensure the "information availability" required to conduct vital
business operations.
• Timeliness: Defines the exact moment or the time window (a particular time of the day, week, month, and year)
during which information must be accessible. For example, if online access to an application is required between
8:00 a.m. and 10:00 p.m. each day, any disruptions to data availability outside of this time slot are not
considered to affect timeliness.
6 Huawei Confidential
Key Concepts — RPO and RTO
Key indicators for measuring the DR system
Week Day Hour Min Sec Sec Min Hour Day Week
RPO RTO
Switch
time
1 Introduction to Business
Continuity
2 Introduction to Backup Solution
8 Huawei Confidential
The Challenges for Data Protection
9 Huawei Confidential
Basic Concepts of Backup
Definition
• A backup is an additional copy of production data, created and
retained for the sole purpose of recovering lost or corrupted data.
Organizations must find an efficient, cost-effective backup method
for ever-increasing amounts of data that meets business and
regulatory demands for data storage, retention, and availability.
Database HDDs
Operation system
Optical discs
Application Backup
Tapes
Document
… …
10 Huawei Confidential
Basic Knowledge of Backup — Comparison
Backup vs Disaster Recovery vs Archive
Backup Disaster Recovery
Data loss can be resolved Data loss cannot be resolved Backup focuses on
Data corruption can be resolved Data corruption cannot be data recoverability,
Historical data can be retrieved
resolved while disaster
Can be restored to the nearest recovery focuses on
Cannot be directly taken over.
running time business continuity.
Can be directly taken over.
Backup Archive
Used to restore data due to Used to store information for a
Backup is used for
damage or loss long time.
data restoration, and
SLA compliance (RTO/RPO) Meet regulatory compliance and archiving is used for
enterprise efficiency regulatory compliance.
The original data is retained in
requirements.
the original location
Retain only data copies and
delete original data.
11 Huawei Confidential
Basic Knowledge of Backup — Comparison
Comparison of data protection methods
RAID Replication Snapshot Backup
Physical error Natural disaster Yes Yes
(47%) Hardware failure Yes Yes Yes Yes
Human error Yes Yes
Logical error
Software failure Yes Yes
(53%)
Virus Yes Yes
Recovery point Seconds,
Acceptable amount of lost data No data loss Hours Days
objective (RPO) minutes, or hours
Recovery time
Data recovery period Real-time Minutes or hours Minutes Hours or days
objective (RTO)
Several weeks,
How far back can the data be Cannot be Cannot be Several hours
Data retention period recovered to recovered recovered or days
months, or
years
The objectives of data protection are to prevent physical and logical errors and meet
the requirements of RPO, RTO, and retention length.
No single data protection algorithm can serve all these purposes. Therefore, a
complete data protection system consists of multiple algorithms.
12 Huawei Confidential
Backup Strategy — Component
A backup system commonly uses the client-server architecture with a backup server and
multiple backup clients.
Backup/ Backup server Media server
Application client
Functions Functions
Functions • The backup
CommCell
Index • Media server is
• The role of a server manages responsible for
Console Cache
backup client the backup MA writing the data to
is to gather CS operations and the backup device.
Application the data that maintains the • In many cases, the
data is to be backed Metadata backup catalog, media server is
up and send it which contains integrated with the
to the storage information backup server, and
node. about the backup both are hosted on
• It sends the configuration and Data storage the same physical
tracking backup metadata. platform.
information to Security & Licensing • Backup metadata
File system • A backup device is
the backup contains attached directly or
data
server. information through a network
about the to the storage
backed-up data. node's host platform.
13 Huawei Confidential
Basic Knowledge of Backup — Backup Operations
When a backup operation is initiated, significant network communication takes place between the
different components of a backup infrastructure. The backup operation is typically initiated by a
server, but it can also be initiated by a client.
Application server/ 1 Backup server initiates scheduled backup process.
Backup clients
2 Backup server retrieves backup-related information
from backup catalog.
File
server
Database Mail
server server
3 Backup server instructs storage node to load backup
media in backup device.
Application server/
Backup clients
Full backup is a backup of the Cumulative backup copies the Incremental backup copies the
complete data on the data that has changed since the data that has changed since the
production volumes. It provides last full backup. This method last full or incremental backup.
a faster recovery but requires takes longer than an The volume of data backed up is
more storage space and also incremental backup but is faster restricted to the changed data
takes more time to back up to restore. only.
data.
16 Huawei Confidential
Backup Strategy — Topology: LAN Base
Description Production center Backup center
LAN
Backup data flows are transferred
over a LAN. ...
MA CS
17 Huawei Confidential
Backup Strategy — Topology: LAN-Free
Description Production center Backup center
LAN
Backup data flows are transferred
over a SAN. ...
MA CS
Advantages Application Database File Application Backup Backup
server server server server media server
Minor impact on the existing service server
SAN
network
Excellent backup performance
18 Huawei Confidential
Backup Strategy — Topology: Server-Free
Description Production center Backup center
Backup data flows are transferred LAN
19 Huawei Confidential
Backup Topology Comparison
Backup Topology Advantage Disadvantage
• Backup data flows affect network
• Take full use of the existing LAN
LAN Base • Flexible scaling
performance.
20 Huawei Confidential
Backup Strategy — Data Deduplication
Traditional backup processes back up a lot of duplicate data. Backing up duplicate data
significantly increases the backup window requirements and results in unnecessary consumption of
resources, such as storage space and network bandwidth.
Data deduplication is the process of identifying and eliminating redundant data. When duplicate
data is detected during backup, the data is discarded and only the pointer is created to refer the
copy of the data that is already backed up.
File-level deduplication detects and removes redundant copies of identical files.
----It is simple and fast, but does not address the problem of duplicate content inside the files.
Sub-file deduplication breaks files into smaller chunks and then uses a specialized
algorithm to detect redundant data within and across files.
• Fixed-length block deduplication
----Using hash algorithm for block boundary, but this may miss many opportunities to discover
redundant data.
• Variable-length segment deduplication
----Complex but effective
21 Huawei Confidential
Source-Based Data Deduplication
Source-based data deduplication eliminates redundant data at the source before it transmits
to the backup device. Source-based data deduplication can dramatically reduce the amount of
backup data sent over the network during backup processes. It provides the benefits of a shorter
backup window and requires less network bandwidth. There is also a substantial reduction in the
capacity required to store the backup images.
Source-based deduplication increases the overhead on the backup client, which impacts the
performance of the backup and application running on the client. Source-based deduplication
might also require a change of backup software if it is not supported by backup software.
22 Huawei Confidential
Target-Based Data Deduplication
Target-based data deduplication runs on the backup device, which offloads the backup client
from the deduplication process. The backup client sends the data to the backup device and the
data is deduplicated at the backup device, either immediately (inline) or at a scheduled time (post-
process).
• Inline deduplication performs deduplication on the backup data before it is stored on the backup
device.
• This method reduces the storage capacity needed for the backup but needs a large backup
window.
• Post-process deduplication enables the backup data to be stored or written on the backup
device first and then deduplicated later.
----This method is suitable for situations with tighter backup windows and large storage capacity.
23 Huawei Confidential
Contents
1 Introduction to Business
Continuity
2 Introduction to Backup Solution
24 Huawei Confidential
What Is Disaster Recovery?
• Disaster recovery (DR) is the coordinated process of restoring systems, data, and the
infrastructure required to support ongoing business operations after a disaster occurs. It is the
process of restoring a previous copy of the data and applying logs or other necessary
processes to that copy to bring it to a known point of consistency. After all recovery efforts
are completed, the data is validated to ensure that it is correct.
25 Huawei Confidential
Service Interruption May Cause Great Losses to
Customers
Hacking
Misoperation Building
Network fire Terrorist
Software problems attacks
Power
faults failures Natural
Device faults
Data disasters
corruption Cooling system
issues
! Service interruption
Disaster recovery (DR) focuses on application continuity. Backup focuses on data recoverability.
26 Huawei Confidential A good BC&DR system must incorporate these capabilities.
Two Indicators About DR Construction
RPO: Recovery Point Objective (amount of lost data caused by downtime) RTO: Recovery Time Objective (downtime)
Tier 6 - Zero data loss and Tier 7 - Zero data loss and
0 < 15 minutes Active-active/DPA
remote cluster support automated service recovery
Tier 5 - Real-time data Tier 6 - Zero data loss 0 < 2 hours Active-passive/DPA
transmission and complete
equipment support Tier 5 - Two-site two-phase
2 to 12 hours < 24 hours Active-passive/DPA
commit
Tier 1 - Basic support Tier 0 - No off-site data Days to? Days to? Backup solution
27 Huawei Confidential
Major Disaster Recovery Technologies
Application Application
Database mgmt.
Database mgmt. system
Host layer system
OS OS File system
File system • Application
replication
Raw device/volume • Database replication Raw device/volume
• Logical volume
Device I/O driver replication Device I/O driver
IP
SAN-based network SAN IP
SAN
layer
28
DR Implementation in the Industry
DR is implemented on hosts without compatibility requirements for underlying devices.
Host layer (typical Pros During database replication, the DR center can take over part of the work from the production
replication software such center.
Database replication can be implemented only for the corresponding database.
as Symantec VVR, Oracle
Data Guard, DSG and
Cons Host-layer replication takes up certain host resources and affects the application system.
DR is implemented on hosts, which is complex and usually requires system reconstruction.
Quest)
Broad compatibility and resource integration of different back-end heterogeneous SAN storage.
Network layer (typical Pros Simultaneous disaster recovery for multiple SAN arrays without one-to-one mapping.
Extendable disaster recovery platform.
No extra investment required as the number of hosts and arrays increases.
gateway devices such as
IBM SVC and EMC VPLEX) Cons High initial investment and few available vendors for such a solution.
Complex gateway networking increases latency.
write or mirroring, such as The storage arrays at both ends must be from the same vendor. Some of them can be heterogeneous (for example,
Cons
Huawei supports heterogeneous virtualization).
Huawei storage series) For some vendors, data cannot be accessed from the remote site in real time (Huawei active-active SAN architecture
supports real-time access). Data can only be viewed through an active/standby switchover or using snapshots.
29
Architecture and Technology
Overview of Huawei DR solutions Remote DR solutions
Intra-city DR solutions
Local HA solutions DR data center solution (geo-
DR data center solution (active- redundant mode)
active mode) Disaster recovery data center
Local High Availability Disaster recovery data center solution (active-passive mode)
(HA) solution solution (active-passive mode)
> 100 km
≤ 100 km
Local HA Solution
Highlights
High availability (HA) for
FusionSphere VM VM VM VM VM VM FusionSphere local applications.
VMware VMware
In a DC, storage is hardened
to eliminate the risks of a
single storage device.
No data is lost, and services
are not interrupted (RPO =
Oracle/SQL server Oracle/SQL server 0, RTO = 0).
This solution can be
smoothly expanded to
SAN
disaster recovery data center
solution (active-passive
Third-party
storage system
Synchronous OceanStor 18000 V5 mode).
mirroring
31 Huawei Confidential
Local HA Active-active Active-passive Geo-redundant
Production center 1
WAN
Production center 2
Highlights
GSLB
The active-active high-reliability service
cluster architecture ensures zero data loss and service
V V V V V V (F5/L280 V V V V V V interruption upon a DC failure (RPO = 0, RTO =
M M M M M M M M M M M M
0) 0).
FusionSphere,
VMware, Two DCs provide services concurrently, service
WebLogic, or loads are balanced, and resources are accessed
WAS cluster based on the proximity access principle.
Oracle, DB2,
Virtualization deployment is implemented, and
or SQL Server middleware clusters are automatically expanded
cluster and scaled down.
Storage, databases, networks, as well as
virtualization and transmission are optimized,
SAN SAN
ensuring the same service experience between
the local site and a site at a distance of 100 km.
The physical active-active topology and logical
HyperMetro
service topology are visually presented, and the
operating status of the active-active DCs is
monitored.
32 Huawei Confidential
Local HA Active-active Active-passive Geo-redundant
34 Huawei Confidential
Thank you. Bring digital to every person, home, and
organization for a fully connected,
intelligent world.