ibm xiv storage system series asynchronous mirroring
TRANSCRIPT
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 1/16
IBM Systems and Technology
Technical White Paper
IBM XIV Storage
System Series Asynchronous Mirroring Promoting System Optimization for Business Continuity
Contents
2 The impact of dataunavailability
4 Key requirements of asynchro-
nous mirroring solutions
6 Powerful XIV mirroring features
9 High performance mirroring
10 Flexible XIV deployment and
implementation
11 Minimal administration
13 Low total cost of ownership
13 Conclusion
14 The XIV Storage System series
Executive summary This paper highlights the importance of mirroring as a data protection
strategy to minimize risk of data and system unavailability and the
advantages of IBM® XIV ® Storage System asynchronous mirroring.It covers the system’s essential mirroring solution characteristics and
organizational mirroring requirements that ensure business continuity.
While this paper touches on the value of synchronous mirroring in
certain situations, its primary focus is asynchronous mirroring.
Reading this paper will give you a clear understanding of effective
enterprise mirroring implementations and the IBM XIV Storage System
series solution, which offers a comprehensive, successful mirroring
strategy for today and the future.
IntroductionEnterprises rely on their data for business success. It is impossible for
most organizations to function day-to-day without online information
about customers and products or the data infrastructure that supports
business processes. Organizations use data replication to ensure data is
available in the event of a system failure or a disaster that degrades system
functionality or renders a data center inaccessible. Regardless of the
cause—equipment failure, natural disaster, or terrorist attack—companies
need to implement a method to minimize and remedy the adverse
business outcomes of these situations, which can heavily impact revenues,
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 2/16
2
IBM Systems and Technology
Technical White Paper
expose organizations to high risk, and cause a slew of other
serious problems. Data mirroring , one of the primary data
replication methods, enables organizations to preserve data
by copying it to secondary systems as it is created online.
Measuring the impact of data
unavailability Two measures define mirroring solution recovery requirements
for organizations: recovery point objective (RPO) and recovery time
objective (RTO). RPO describes the acceptable amount of data
loss measured in time. The RPO is the point in time to which
one must recover data as defined by the organization. This is
generally a definition of what an organization determines is an
“acceptable loss” in a disaster situation.
RTO defines the maximum acceptable amount of time torestore both data and the systems to access the data in order
to prevent unacceptable consequences associated with a break
in business continuity.
The optimal RPO and RTO are zero (zero data loss and zero
recovery time following a disaster), but such objectives may
not be always possible—or justified—due to implementation,
support, and recovery costs, and the latency resulting from the
distance between replication sites.
While mission-critical applications typically require a very low
RPO, non-critical applications may be able to suffice with ahigh RPO. For instance, a backup routine that runs once a
day can have an RPO of 24 hours. Assigning a low RPO to
non-critical applications wastes precious system and network
resources.
Acceptable
data loss
0+ Seconds Minutes Hours
Critical applications;
financial
data,transactions
Low priority
backup applications;
testingenvironment
Productivity applications;
RPO
Figure 1: RPO and data value. The higher the RPO, the higher the potentialdata loss
Synchronous versus asynchronous mirroring
Synchronous mirroring and asynchronous mirroring are two
approaches for accommodating application-mirroring
requirements.
With synchronous mirroring, data written by a host to a storage
volume (the Master peer on a primary system) is replicated to
a secondary storage system (the Slave peer) before the write
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 3/16
3
IBM Systems and Technology
Technical White Paper
Synchronous mirroring
Asynchronous mirroring
Host Master Slave
Host Master Slave
1. Host writes new data to Master
4. Master acknowledges write by host
Extended response time lag
Synchronization gap between peers2. Master acknowledges write by host
1. Host writes new data to Master
2. Master replicates to Slave
3. Master replicates to Slave
3. Slave acknowledges write by Master
4. Slave acknowledges write by Master
Figure 2: Synchronous and asynchronous mirroring. With synchronousmirroring, RPO is 0, yet response times increase as the distance between
peers increases. With asynchronous mirroring response times are notdependent on link bandwidth and quality, but the RPO is greater than 0.
completion acknowledgement is sent back to the host. This
approach yields complete data currency between the Master
and the Slave (RPO of 0), yet introduces latency while the
host waits for acknowledgement.
With asynchronous mirroring, data written to the Master
is acknowledged immediately to the originating host, yet is
replicated to the Slave at a later time. Therefore, asynchronous
mirroring offers lower write latency at the cost of increased
RPO (less data currency and lack of synchronization between
the Master and Slave).
Asynchronous mirroring is advantageous for a variety of use
cases and is a compelling solution for replication between
distant sites, since it eliminates the latency inherent to
synchronous mirroring; it can offer lower implementation
costs as well.
Effective asynchronous-mirroring planning and implementation
can both minimize the currency gap between mirrored peers
and help organizations realize better data availability with lower
implementation and operational costs.
The RPO possible with asynchronous mirroring is a function
of variables such as network bandwidth, line quality, distance,
latency, and mirroring system functionality. RPO can range
from mere seconds to minutes or hours, depending on
application requirements and the mirroring implementation. With its low latency and flexibility in supporting low and
high RPOs, asynchronous mirroring is an attractive option,
particularly when long distance, low latency replication is
required.
IBM XIV Storage System series asynchronous mirroring
The IBM XIV Storage System series, with pioneering
architecture and a powerful feature set, offers comprehensive
asynchronous mirroring functionality that helps guarantee data
availability in the event of failure or disaster, providing valuable
enterprise-class benefits not available in traditional mirroringsolutions.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 4/16
4
IBM Systems and Technology
Technical White Paper
Key requirements of asynchronous
mirroring solutionsIn addition to being reliable, mirroring solutions need to
provide other essential data replication features:
Requirement #1: An enterprise-class
feature setLow RPO for prioritized applications. Because of the severe
repercussions that data loss can have on organizations, minimal
data loss with quick data recovery is a principal mirroring
solution requirement.
Synchronous mirroring enables administrators to restore all
Master data from its corresponding Slave peer at the price of
increased write latency. This solution, however, is appropriate
for replication distances no greater than 62 miles/100 km(potentially up to 186 miles/300 km—with more powerful
equipment and infrastructure). Synchronous mirroring is less
appropriate over links where latency is already pronounced.
Asynchronous mirroring virtually eliminates latency concerns at
the price of reduced data currency, with the RPO reflecting
the organization’s acceptable amount of data loss in time units.
An RPO setting of 30 seconds indicates that the Slave’s data
currency should be no more than 30 seconds older than the
Master’s. An RPO equal to or less than 30 seconds is common
for many high-end mirrored applications.
Multiple RPOs. In an optimal environment, all applications
would mirror with an RPO of 0. Factors governing application
profiles, disaster recovery (DR) requirements, replication
topology, infrastructure, and cost rarely enable organizations
to realize this objective.
Assigning the same RPO for all mirrored applications regard-
less of their data’s criticality can typically result in suboptimized
bandwidth utilization and impaired performance. In order to
balance application criticality with available resources and net-
working bandwidth, most organizations prefer instead to set
individual RPOs per application. Assigning application-based
RPOs can help organizations optimize bandwidth utilization
for mirroring, accommodating a mix of low and high criticality
applications.
Support for application-consistent recovery points. An
application-consistent recovery point represents the data whose state
is guaranteed to be consistent from an application perspective.
Creating an application-consistent snapshot on a storage system
requires integrated host software that generates the snapshot
only after ensuring an application-consistent state. In addition,
creating application-consistent snapshots requires host process-
ing, which can adversely affect application response time.
A crash consistent recovery point recovers data to the point in time
at which the crash occurred, but does not guarantee application
data integrity. Therefore, to promote recovery data integrity,
an application-consistent recovery point is superior to a crash
consistent recovery point.
The ability to support both application- and crash-consistent
recovery for any mirror is extremely advantageous, since it
allows organizations to implement their preferred recovery approach based on data criticality.
Ability to consistently mirror a group of volumes.
Applications may require consistent replication of more than
one volume. For example, a database application that spans
multiple volumes requires cross-volume mirroring consistency.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 5/16
5
IBM Systems and Technology
Technical White Paper
Requirement #2: Enterprise-class
mirroring performanceHigh-performance mirroring is critical for facilitating low
RPO replication with little to no impact on general application
performance. The storage system should provide high perform-
ance even when mirroring is configured over suboptimal links
that experience intermittent disruptions.
Requirement #3: Flexible mirroring
deployment and implementationFlexible topology configurations. The most common
mirroring topology is a one-to-one mirror between systems
at different sites, with multisite organizations increasingly
implementing—or considering—multisystem/multisite
mirroring configurations. Two popular options for such
deployments are the many-to-one and one to-many topologies(see Figures 7 and 8 below).
With the one-to-many configuration, a single system (hosting
one or more Master volumes) replicates to several systems
(functioning as Slaves). Each of the Master’s mirrored volumes
replicates to one of several Slaves. This enables the organization
to distribute replication load across multiple links and share the
recovery responsibility among multiple sites.
A many-to-one configuration replicates multiple independent
Masters to a single designated Slave system at a single recovery
site. As with the one-to-many configuration, the many-to-one
configuration also leverages multiple links for distributed
replication.
A hybrid option (see Figure 9 below) combining characteristics
of both one-to-many and many-to-one topologies is not
uncommon.
Organizations are increasingly considering using iSCSI
(Internet Small Computer System Interface) links for low-cost
inter-site mirroring. Customers expect advanced mirroring
solutions to offer mirroring over both Fibre Channel (FC)
and iSCSI.
Bidirectional mirroring . Traditional storage mirroring imple-
mentations may be unidirectional—with replication from a local
system (also referred to as a Primary system) to a remote system
(also referred to as a Secondary system)—but not the other way
around. More advanced configurations allow each system to
function as a Master for certain volumes and as a Slave for other
volumes. This configuration enables organizations to distribute
the application load and traffic to better accommodate business
needs and performance goals offering greater storage deploy-
ment flexibility.
Ability to accommodate changing mirroring requirements.
Mirroring requirements are dynamic; they are dependent upon
business requirements, infrastructure changes, and/or new
application priorities. Increasingly, organizations are evaluating
improved disaster recovery topologies for their environments,
wanting to make changes to their topologies with minimal
effort. For example, an organization may want to change an
existing synchronous mirror configuration to asynchronous
replication. Being able to make this type of change without
reinitializing the Slave helps minimize mirroring setup time,
maximizing system availability and optimizing utilization of network bandwidth for other mirror processing.
Requirement #4: Minimal administrative
overhead Mirroring deployments are notorious for their complexity and
heavy setup and implementation requirements—as well as hefty
planning, training, and setup time. Overly imposing and cryptic
mirroring documentation often makes the implementation
more cumbersome.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 6/16
6
IBM Systems and Technology
Technical White Paper
Mirroring complexity notwithstanding, enterprise-class mirror-
ing should be streamlined, offering significantly lower effort
than traditional solutions. The enterprise mirroring solution
should feature:
1. Easy and quick setup. Mirroring configuration options
should be laid out clearly, with remote connectivity easily
defined through a graphical user interface (GUI) or a
straightforward command line interface (CLI). Volume
mirroring should be a seamless process, with minimum
setup steps.
2. Streamlined management . Mirroring management should
feature streamlined status monitoring with automatic
alerting. Mirroring displays should be simple and straight-
forward. Common mirroring management tasks, such
as changing a mirroring role or adding a volume to a
consistency group, should be easy to perform.
3. Simple testing and failover processes . The mirroring solution
should enable customers to easily test and validate the
mirroring configuration and replica. More importantly,
the solution must feature a reliable, no-fuss failover
process that does not mandate extensive training.
4. Comprehensive and clear reporting and alerting . The mirror-
ing solution should feature effective, helpful mirroring
status reporting and alerts when the RPO cannot beaccommodated, and clearly indicate the networking issues
that risk attainment of the RPO.
The powerful XIV storage mirroring
feature set The IBM XIV storage meets and exceeds enterprise-class
mirroring requirements through a rich set of features.
Proven snapshot-based technology for asynchronous
mirroring . Asynchronous mirroring employs patented
IBM XIV breakthrough snapshot technology, which streamlines
implementation while demonstrating stellar performance for
low-RPO mirroring applications. In contrast with traditional
snapshot implementations, IBM XIV snapshot technology is
an inherent part of the base system architecture; it has not been
appended to a preexisting system architecture. This level of
integration obviates architectural restrictions that have tradi-
tionally prevented systems from providing high snapshot related
performance.
Notable IBM XIV snapshot mirroring features include:
● Support for a large number of snapshots● Functionality to snapshot a volume, a volume consistency
group, or another snapshot ● Ability to restore a volume, a volume consistency group,
or a snapshot—from a snapshot ● High performance by leveraging effective snapshot
technology
Smart IBM XIV asynchronous mirroring design . With
asynchronous mirroring, mirror-related snapshots are used
to determine the data that needs to be replicated. The system
uses these snapshots to capture the state of the Master being
mirrored.
XIV asynchronous mirroring attains synchronization between
Master and Slave peers through a recurring data replication
process, called a Sync Job, which is triggered automatically according to user-configurable schedules. (The Sync Job can
also be triggered manually or via a script.) The Sync Job takes a
snapshot of the Master and looks for data differences between
that snapshot and the previous mirroring-related Master snap-
shot on the Slave. The Sync Job then replicates Master data
corresponding to these differences with the Slave. At the
completion of a Sync Job, new mirroring-related snapshots
are created on the Slave and the Master.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 7/16
7
IBM Systems and Technology
Technical White Paper
Low RPO for critical applications. The IBM XIV system
supports mirroring RPOs as low as 30 seconds and
replication intervals as low as 20 seconds.
Independent RPO and interval. In addition to IBM XIV
system support for broad RPO and interval ranges,
the IBM XIV mirroring implementation allows administrators
to set RPOs and intervals for each mirror independently
(or per IBM XIV consistency group). This enables administra-
tors to accommodate mirroring requirements based on applica-
tion criticality, rather than forcing them to set the same RPO
for all applications—which could cause inefficient utilization
of system resources and adversely impact performance.
Automatic and manual replication . The IBM XIV system
supports both automatic and manual initiation of
asynchronous replication jobs. With automatic replication,
the system allows administrators to specify schedules based
on intervals ranging from 20 seconds to 12 hours. Manual
replication makes it possible to create an ad hoc synchroniza-
tion job or to trigger a synchronization job from a script.
Manual replication can be coordinated with host activitiesto create snapshots that represent application-consistent
recovery points.
For additional business continuity benefit, the system allows
administrators to concurrently employ both manual and
schedule-based replication types for each mirror—enabling
manual synchronization for mirrors already using an interval
based schedule.
Support for both application- and crash-consistent
recovery . The XIV system supports both application-
and crash-consistent recovery points for any given mirror.
Support for mirroring volume groups. With IBM XIV
system support for mirrored consistency groups, administrators
can mirror volume groups consistently, easily adding volumes
to mirroring groups and safely removing volumes from them
without stopping or otherwise affecting volume mirroring.
Online and off line initialization options. The objective of
the initialization process is to establish an initial replica of the
Master state on the Slave. XIV storage initialization establishes
a baseline for the master state when mirroring begins. IBM XIV
online initialization enables over-the-wire Slave initialization.
Local
site
Master peer Slave peer
most_recent
last_replicated
Replication
last_replicated
Remote
site
Figure 3: Asynchronous mirroring snapshots. The IBM XIV system maintains three snapshots per mirror coupling: two on the Master and one on the Slave.
The system captures a consistent (recoverable) Master state in the last_replicated snapshot (stored on both Master and Slave). The most_recent snapshotrepresents the most recent Master state that needs to be replicated to the Slave. The system determines the data to replicate by comparing the Master’s
most_recent to the last_replicated snapshots.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 8/16
8
IBM Systems and Technology
Technical White Paper
Online (over-the-wire)initialization
Initialization starts
Initializationreplicates initial
image to the Slave
Master Slave
Master finishes Initialization
Time
Initializationstarts
Initialization finished and
ongoingmirroring can
commence
Figure 4: Online initialization. After a new mirror is defined, a snapshot of the
Master is taken, representing the initial peer state. When mirroring is first acti-vated, the system identifies the data that needs be replicated as part of theinitialization phase. An over the-wire synchronization process then replicates
the data from the Master to the Slave.
The IBM XIV offline initialization feature enables administrators
to initialize the Slave peer without requiring an inter-site link
to replicate the Master to the Slave. When a mirror is first acti-
vated after offline initialization, the system takes a snapshot of the Master, compares that snapshot with the Slave replica, and
replicates differences to the Slave. One of the salient benefits
of an off-line initialization implementation is that prior to
ongoing mirroring, the system verifies the Slave’s contents
against the Master and does not automatically assume the
Slave is a valid replica.
Offline initialization
Initialization starts
Initialization comparesimages and only
replicates differences
Initialization complete
Initialization
starts
Master Slave
Time
Initialization
finished and
ongoing
mirroring can
commence
Figure 5 : Of fline initialization. Instead of running an over-the-wire initializa-
tion, a backup of the Master can be shipped to a remote recovery site. Oncethe backup is restored on the Slave, the system can establish mirroring afterfirst comparing the Master with the restored Slave replica. This approach
can shorten the initialization process considerably and optimize bandwidthutilization. The mirror is defined with a parameter indicating that a Masterreplica exists on a remote system. Based on this parameter, the system
determines that offline initialization is set.
Maximized data resilience thanks to mirroring . With
mirroring, an IBM XIV system can overcome a storage
media error on the Master system using data from its Slave.
IBM XIV supports both synchronous mirroring and
asynchronous mirroring recovery (as long as the media
error is in a partition shared between the volume and the
Last Replicated Snapshot ).
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 9/16
9
IBM Systems and Technology
Technical White Paper
IBM XIV mirroring also enables effective scrubbing that com-
pares data between local and remote systems. IBM XIV scrub-
bing is a nonstop background activity that verifies the integrity
and redundancy of stored data by examining each data partition
on the system. This process enables essential early detection of
drive read/write errors, minimizing their impact and expediting
recovery processes. The scrubbing process runs continuously
on all modules, checking all drives in parallel, and all drive areas
(including areas never written to). With IBM XIV mirroring,
when the scrubbing process encounters a partition mirrored on
a remote target, it compares the partition copies and looks for
data inconsistencies between primary and remote machines.
High performance with IBM XIV
mirroring
Enterprise-class data replication demands a powerful solution toaccommodate low RPO mirroring requirements without impairing
overall application performance.
Following are notable IBM XIV features that facilitate high
mirroring performance:
High memory utilization minimizes disk access and
maximizes mirroring speed. The IBM XIV system maximizes
the utilization of cache memory for asynchronous mirroring
replication. This minimizes time- and process-intensive disk
access for high performance data replication.
Notably, the IBM XIV caching implementation and large
memory also enable better resilience during link disconnections
and link bandwidth fluctuations, and minimize the mirroring
performance impact on non-mirrored volumes that might
experience cache conflicts.
Low replication granularity minimizes replication traffic.
IBM XIV asynchronous mirroring is extremely efficient, repli-
cating only actual data changes; if one block of data changes,
only one block is replicated. This low granularity replication
helps maximize the network bandwidth available for mirroring
applications.
Independent RPO and interval configuration helps admin-
istrators match RPO with application criticality . IBM XIV
system support for independently set RPOs and intervals
per mirror enables administrators to optimally accommodate
mirroring requirements based on application criticality. One
example is setting a low RPO for critical applications, setting
relatively high RPOs for lower priority applications—and
specifying the same minimal interval for all mirrors. This con-
figuration permits the system to prioritize replication based on
application RPO—which helps promote increased performance.
Offline initialization maximizes bandwidth and reduces
initialization time. Offline initialization reduces mirroring
initialization time, maximizing available network bandwidth for
non-initialization replication traffic. The benefit is pronounced
with low-bandwidth mirroring links.
Manual replication helps optimize mirroring scheduling .
IBM XIV manual replication makes it possible for administra-
tors to schedule mirroring at specified hours for any given
application to efficiently accommodate traffic peaks andmaximize the replication rate.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 10/16
10
IBM Systems and Technology
Technical White Paper
Multiple mirroring target support maximizes potential
replication bandwidth . IBM XIV system support
for multiple target connectivity enables organizations to
harness expanded connectivity bandwidth for mirroring
instead of requiring all mirroring traffic to go through a
single dedicated line.
Replication throttling enables control over synchronization
rates per target . The IBM XIV mirroring solution features the
ability to dynamically accommodate connections with fluctuat-
ing bandwidth and latency. The system examines the bandwidth
and I/O latency of the mirroring link and adjusts its I/O rate
to the estimated available bandwidth. This accommodation of
fluctuating bandwidth characteristics minimizes potential repli-
cation disruptions that could degrade performance.
Flexible IBM XIV deployment and
implementation IBM XIV asynchronous mirroring presents administrators with
flexible deployment and implementation options, streamlining both
setup and maintenance activities alike.
Connectivity with multiple targets allows for flexible
deployment choices● Each IBM XIV system can set mirroring relationships with
up to eight other systems● Each IBM XIV system features simultaneous support for
separate replication sources and targets per system. A systemcan therefore function as a Master peer for some mirrors and
as a Slave peer for other mirrors
● Each IBM XIV system can support hundreds of mirrors● Each mirror can be configured to run independently over
FC or iSCSI● A mirror can be defined either for a single volume or for a
volume group (consistency group)● The system supports both synchronous and asynchronous
mirroring● Each mirror can have an independent schedule type and
interval● Mirror topology configurations supported include:
one-to-one, one-to-many, many-to-one, and hybrid
XIV System2 XIV System3
XIV System4 XIV System5
XIV System6 XIV System7
XIV System8 XIV System9
XIV Sys tem1
Figure 6 : Each system can set mirroring relationships with up to 8 othersystems
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 11/16
11
IBM Systems and Technology
Technical White Paper
Figure 7 : Many-to-one topology. One system (or site) serves as a centralhub target for all other systems
Figure 8: One-to-many topology. One system (or site) replicates to multiple
targets
Figure 9: Hybrid IBM XIV topology. Some systems (or sites) function as bothMaster and Slave
Independent RPO and interval configuration helps
administrators tailor RPO to application priority and
available resources. IBM XIV system support for independ-
ently set RPO and interval per mirror enables administrators
to optimally accommodate mirroring requirements based on
application criticality.
Minimal administration with IBM XIV
asynchronous mirroring Asynchronous mirroring is simple to set up and manage, with
straightforward support for mirroring failover and failback.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 12/16
12
IBM Systems and Technology
Technical White Paper
Powerful GUI functionality . The IBM XIV storage sports an
intuitive GUI console and a simple command line interface that
facilitate streamlined administration, supporting these key mir-
roring features:
● Clear status monitoring and alerting: the system generates
system events and monitors critical asynchronous mirroring-
related processes to produce important mirroring assessment
information and statistics● Easy mirroring schedule assignment and adjustment ● Ability to easily change mirroring peer roles from Master
to Slave and vice versa● Straightforward failover and failback ● Simple deployment configuration using graphical
topology tools
Figure 10: The IBM XIV GUI displays a user-friendly view of the deploymenttopology
Independent RPO and interval settings. IBM XIV adminis-
trators can specify an independent RPO per mirror without
adjusting the mirroring interval. Assignment of the same
minimal interval to all mirrors enables the system to prioritize
replication with no user intervention.
Consistency group support . IBM XIV administrators can
easily manage a mirrored volume group using commands that
apply to the whole group, including activation, deactivation,
role change, and streamlined restoration of consistent volume
groups from a remote site after local site disruption.
Automatic dynamic sync rate control. With IBM XIV
dynamic sync rate control, no administrator intervention
is needed to tune the system to accommodate fluctuating
link speed.
Easy failover process. The IBM XIV system facilitates easy
mirroring failover through a single command that changes the
role of the specified peer.
Easy support for planned maintenance scenarios. The
IBM XIV system makes it easy to validate the mirror replica
on the Slave without disrupting the mirroring process. The
administrator simply creates a duplicate of the mirrored replica
and maps that copy to a host.
When both mirroring peers are available, IBM XIV facilitateschanging peer roles by enabling mirror switching. With this
feature, the administrator instructs the system to simultaneously
switch the roles of both peers in a specified mirroring relation-
ship; the Master changes to become the Slave and vice versa.
Quick, streamlined failback . The system facilitates a stream-
lined failback process by supporting offline initialization.
The storage administrator initializes the original Master with
data from the current Master (the promoted Slave). As already
noted, offline initialization helps an organization avoid long
resynchronization periods, especially with suboptimal lines,
and realize improved bandwidth utilization.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 13/16
13
IBM Systems and Technology
Technical White Paper
Figure 11: The IBM XIV mirroring GUI features a clear status display andeasy-to-use functionality, such as change role.
Low total cost of ownershipThe IBM XIV asynchronous mirroring function can help organiza-
tions realize reduced setup and operational costs—which translates
into lower TCO.
Offline initialization helps save precious time and cost .
With offline initialization, administrators can expedite initial-ization by eliminating over-the-wire traffic. This approach
benefits performance and leads to cost savings as well. A good
case for offline initialization is with mirroring failback, in which
restoring a Master after failover can demand long, costly initial-
ization; offline initialization eliminates this overhead. Offline
initialization can also avoid over-the-wire traffic when changing
mirroring type from synchronous to asynchronous.
Support for mirroring over iSCSI helps reduce costs. Due
to its potential cost savings, replication over iSCSI is becoming
a viable mirroring option for enterprises. The IBM XIV system
supports mirroring over iSCSI, facilitating streamlined deploy-
ment at reduced costs.
Mirroring leverages thin-provisioning functionality .
The IBM XIV system supports mirroring of thinly provisioned
volumes, ensuring efficient, low-cost space consumption and
allocation for mirroring applications.
Mirroring is an integral function of the IBM XIV solution
at no extra cost . With the IBM XIV storage system series, all
mirroring functionality and features are available out-of-the-box
at no extra charge:
● Synchronous mirroring● Asynchronous mirroring● Scheduled and manual replication● Support for mirrored volumes and consistency groups● Mirroring failover and recovery ● Online and offline initialization● Graphical user interface for storage management
Effective bandwidth utilization . With IBM XIV asynchro-
nous mirroring, only the data changes on the Master peer
replicate to the Slave peer: if one block changes, only one block
is replicated. The IBM XIV system maintains strong control
over synchronization rates per mirror, which helps attain
efficient bandwidth utilization.
Optimized performance warrants no extra hardware. High
performance asynchronous mirroring does not warrant special
hardware or hardware upgrades, e.g., cache upgrades. The
IBM XIV system maintains the optimal hardware configuration
to support enterprise-class mirroring.
Conclusion
Data is often considered the most important asset of an enter-prise and absolutely critical to business success. The IBM XIV
system features powerful asynchronous mirroring that
satisfies short- and long-term enterprise-class requirements
for data mirroring. With the IBM XIV system, asynchronous
mirroring is easy to set up and offers flexible configuration
options, providing simple data restoration and recovery to
support organizations’ business continuity objectives. The
special blend of breakthrough IBM XIV storage architecture
and features places its asynchronous mirroring solution in
the highest echelon of mirroring solutions by ensuring data
availability in the event of failure or disaster and maintaining
enterprise-class performance with low RPO/quick RTOmirroring, while minimizing overall costs.
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 14/16
14
IBM Systems and Technology
Technical White Paper
About the IBM XIV Storage System series
IBM XIV storage is proven, high-end disk storage designed
by listening to customers and addressing their storage chal-lenges across the broadest spectrum of business applica-
tions. The XIV series of fers highly affordable storage
suitable for even the most demanding workloads, providing
tier 1 consistent high performance and high reliability at tier
2 costs. Never compromising performance for reliability, the
XIV grid architecture deploys massive parallelism to allocate
system resources evenly at all times and scale seamlessly,
without the need for complex, time-consuming tuning and
configuration. Comprehensive XIV storage disaster recovery
infrastructure and functionality, including synchronous and
asynchronous mirroring, self-healing, internal data redun-
dancy, and flexible failover and failback scenarios promote
reliable, effective, expedited enterprise disaster recoverystrategies.
A recognized industry leader in enterprise storage
manageability, the XIV system sets a new standard for
ease of use by automating most tasks and providing an
exceptionally intuitive user interface. It also offers “any-time,
anywhere” monitoring via the IBM XIV Mobile Dashboard
app for mobile devices. The XIV architecture enables
performance to grow with capacity and tightly meshes
with cloud technologies for even greater agility in meeting
business needs.
The XIV Storage System has met with rapid marketacceptance and success, with thousands of installations in
diverse industries worldwide, including financial services,
healthcare, energy, education and manufacturing. The
XIV series supports a wide range of workload needs, from
capacity-hungry to ultrahigh per formance. It integrates
easily with virtualization, email, database, analytics, data
protection and other solutions from leading providers,
such as Microsoft, IBM, SAP, Oracle, SAS, VMware
and Symantec. The XIV series plays a key role in
IBM’s end-to-end dynamic infrastructure solutions,
integrating with IBM ProtecTIER®, SONAS, SAN Volume
Controller, IBM Storwize® V7000 and IBM Tivoli® products.
For more information To learn more about the IBM XIV Storage System series,
please contact your IBM representative or IBM Business
Partner, or visit the following website:
ibm.com /storage/disk/xiv
Additionally, IBM Global Financing can help you acquire the
IT solutions that your business needs in the most cost-effective
and strategic way possible. We’ll partner with credit-qualified
clients to customize an IT financing solution to suit your busi-
ness goals, enable effective cash management, and improve your
total cost of ownership. IBM Global Financing is your smartest
choice to fund critical IT investments and propel your business
forward. For more information, visit: ibm.com /financing
About the author Rami Elron is a senior product manager within the
IBM Systems and Technology Group. He is a subject
matter expert in storage architecture, mirroring, and security.
He has been helping drive technical strategy and innovation
for the IBM XIV Storage System series since 2008. He can be
contacted at [email protected].
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 15/16
Notes
7/30/2019 IBM XIV Storage System Series Asynchronous Mirroring
http://slidepdf.com/reader/full/ibm-xiv-storage-system-series-asynchronous-mirroring 16/16
Please Recycle
© Copyright IBM Corporation 2012
IBM CorporationRoute 100Somers, NY 10589
Produced in the United States of America January 2012
IBM, the IBM logo, ibm.com, ProtecTIER, Storwize, Tivoli, and XIV are trademarks of International Business Machines Corporation in theUnited States, other countries or both. If these and other IBM trademarkedterms are marked on their first occurrence in this information with atrademark symbol (® or ™), these symbols indicate U.S. registered orcommon law trademarks owned by IBM at the time this information
was published. Such trademarks may also be registered or commonlaw trademarks in other countries. Other product, company or servicenames may be trademarks or service marks of others. A current list of IBM trademarks is available on the web at “Copyright and trademark information” at ibm.com /legal/copytrade.shtml
Microsoft is a trademark of Microsoft Corporation in the United States,other countries, or both.
This document is current as of the initial date of publication and may bechanged by IBM at any time. Not all offerings are available in every country in which IBM operates.
The performance data discussed herein is presented as derived underspecific operating conditions. Actual results may vary. It is the user’sresponsibility to evaluate and verify the operation of any other productsor programs with IBM products and programs. THE INFORMATIONIN THIS DOCUMENT IS PROVIDED “AS IS” WITHOUT ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING WITHOUT ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND ANY WARRANTY OR CONDITIONOF NON-INFRINGEMENT. IBM products are warranted according tothe terms and conditions of the agreements under which they are provided.
Statements regarding IBM’s future direction and intent are subject to change or withdrawal without notice, and represent goals andobjectives only.
TSW03094-USEN-00