ESSENTIAL GUIDE TO Replication for · d Storage Essential Guide to Replication for DR ATA REPLICATION as a means of data protection has seen contin- uous and increasing adoption since

1

Managing the information that drives the enterprise

STORAGE

Q

Replication forDR

ESSENTIAL GUIDE TO

INSIDE4 Replication alternatives

14 FAQ: DR replication

18 Data replication tools

22 Disk array-based data replication: Pros and cons

26 Channel Perspective: Benefits of thin replication

30 Vendor resources

Replication for disaster recovery (DR) isno longer a “nice to have” technology,but a necessary part of every business

continuity strategy.

2

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n

nOT TOO LONG AGO, disaster recovery (DR) was an exercise in logistics involvinghundreds of tapes, elaborate transportation schemes and numerouspersonnel assignments. But no matter how well a company prepared,it was almost inevitable that the sheer complexity of the plan meantsomething would fall through the cracks. A single missing tape couldcripple a recovery.

Tape has its place in data protection, including disaster recovery, but not for the mission-critical apps that a companyrelies on to maintain business as usual—or as close to usual as possible. Today’sbusiness environment is almost a cliché:24/7 operations, global reach, interwovennetworks of suppliers and customers, anda sense that a day in the docks could spelldoom for a company. With that kind of urgency, and recovery time objectives(RTOs) that are barely a blink of the eye,tape just doesn’t cut it.

Data replication is the obvious alterna-tive to tape recoveries. Typical replicationsetups called for array-to-array replication,which required duplicating the productionsite configuration at the recovery site. To say the least, duplicating primary storage for recovery purposes often meant an extraordinary expense to ensure business continuity. Other approaches, like host-based replication, are cheaper to implement, but they come at a cost, as well with greater maintenance and administration.

In the last few years, storage managers have received some relief onthe budgetary and administrative sides of the replication equation. Dozensof newer replication products sidestep the constraints of earlier genera-tions of replicators, principally by allowing data to be replicated betweendifferent storage configurations.

On the server side of the shop, virtualization has radically redefineddisaster recovery by breaking the one-to-one, physical server-to-physical

Replication is essential to DR

Once considered a DR solution only for well-heeled enterprises, replication is now a

flexible and affordable alternative for all companies.

Dozens of newerreplication productssidestep the con-straints of earliergenerations of repli-cators, principally byallowing data to bereplicated betweendifferent storageconfigurations.

Copyright 2009, TechTarget. No part of this publication may be transmitted or reproduced in any form, or by any means, without permission in writing from the publisher. For permissions or reprint information, please contact Mike Kelly, VP and Group Publisher ([email protected]).

editorial | rich castagna

mailto:[email protected]

Storage Essential Guide to Replication for DR

server setup that had been required for a recovery site. Now, a singleserver hosting many virtual servers may serve as the DR mechanism tosupport multiple primary servers. And with virtual server setups at bothprimary and recovery sites, the flexibility is almost endless. Fortunately,newer replication products are up to the task and can match the resiliency and flexibility of virtualized server environments.

That’s not to suggest that data replication for disaster recovery isnecessarily a lot easier to implement. There are still a lot of decisions to make, like when to use synchronous vs. asynchronous replication, andthe appropriate types of recovery site storage that won’t bust budgetsbut will still be up to the task of maintaining business processes at anacceptable level.

On the following pages, you’ll get a good look at the various replicationalternatives, where they might be most effectively deployed, and howserver and storage virtualization can be best exploited. Replication maynever entirely eliminate tape from DR scenarios, but it has certainly provided a faster and more reliable alternative. 2

Rich Castagna ([email protected]) is editorial director of theStorage Media Group.

3

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE


dStorage Essential Guide to Replication for DR

ATA REPLICATION as a means of data protection has seen contin-uous and increasing adoption since it first emerged in storagesystems after the first World Trade Center bombing in 1993.Over time, it has evolved into an indispensable component ofdisaster recovery (DR), as well as for operational backup forapplications that require shorter recovery point objectives(RPOs) and recovery time objectives (RTOs) than what tradi-tional tape backups can offer. Firms are also adopting datareplication for remote- and branch-office data protection; in a hub-and-spoke architecture, branch-office data can bereplicated back to central data centers, thus eliminating unwieldy tape-based backup procedures at the branch sites.

The growing adoption of replication services has been drivenby a wide array of data replication products, more lower costreplication offerings, faster and less-expensive networks, andan overall maturing of the technology itself. “Replication-based data protection is among the top three priorities of60% of our clients, which is very different from only a fewyears ago,” said Tim Bowers, global product manager, storageservices at EDS, a Hewlett-Packard (HP) company.

4

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n

ReplicationALTERNATIVES

Data replication is great for protecting criticaldata and ensuring quick recoveries. Find out

where you should deploy replication: in your array, network or servers.

By Jacob Gsoedl

Storage Essential Guide to Replication for DR5

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

NOT ALL REPLICATION IS EQUALAt a macro level, data replication copies data from one storage locationto one or more other local or remote storage systems. But venture beyondthat basic task and you’ll find that data replication products vary in severalkey aspects:

LOCATION: One of the main differentiators among products is wherereplication occurs. The replication service or software can reside on thestorage array, in the network or on the host (server). Array-based replica-tion has been dominating the replication market up to now.

“We did a recent study that shows that in 2007, 83.7% of worldwiderevenue for storage-based replication was done using array-to-arrayreplication, followed by host-based replication with 11.5% and network-based replication with 4.8%,” said James Baker, research manager, storagesoftware at Framingham, Mass.-based IDC. But according to the samestudy, both host- and network-based replication are catching up. Host-based replication is expected to grow at a compound annual growth rate(CAGR) of 18.2% until 2012, while a CAGR of 15.4% is anticipated for net-work-based replication. Both are expected to expand significantly fasterthan the 10% forecasted annual growth for array-based replication.

MODE: Replication can occur synchronously, where data is written tothe primary and secondary storage systems simultaneously; or it can be performed asynchronously, where data is replicated to replicationtargets with a delay. In synchronous replication, the primary storagesystem only commits I/O writes after the replication target acknowledgesthat data has been written successfully. Synchronous replication dependson sufficient bandwidth and low latency, and supported replication dis-tances range from 50 km to 300 km. It’s typically used in apps wherezero RPOs and RTOs are required, such as high-availability clusters andmission-critical applications that demand 100% synchronicity between theprimary and target systems. Conversely, asynchronous replication writesdata to the primary array first and, depending on the implementationapproach, commits data to be replicated to memory or a disk-basedjournal. It then copies the data in real-time or at scheduled intervals to replication targets. Unlike synchronous replication, it’s designed towork over long distances and greatly reduces bandwidth requirements.While the majority of array- and network-based replication productssupport both synchronous and asynchronous replication, host-basedreplication offerings usually only come with asynchronous replication.

TYPE: Replication products can replicate blocks of data on volumes orlogical unit numbers (LUNs), or replication can be performed at the filelevel. With the exception of network-attached storage (NAS), which cansupport both block- and file-based replication, array-based replicationproducts usually operate at the block level. The same is true for network-based replication products. In contrast, most host-based replication offer-


ings operate at the file-system level. Block-based replication isplatform-agnostic and will workseamlessly across various OSes.File-based replication products are very much OS-specific and themajority of available host-basedreplication products are writtenfor Windows. Unlike file-basedreplication, block-based replicationproducts have no knowledge ofthe attached platform, file systemor applications, and depend onauxiliary services like snapshotsfor any type of app integration. Asa result, most storage arrays withreplication support also providesnapshot capabilities that aremore or less integrated with thefile system and key apps like Ex-change and SQL Server databases.

ARRAY-BASED REPLICATIONIn array-based replication, thereplication software runs on one or more storage controllers. It’smost prevalent in medium- andlarge-sized firms, mostly becauselarger firms have deployed higherend storage arrays that come withdata replication features.

With more than 15 years of his-tory, array-based replication is themost mature and proven replicationapproach, and its scalability isonly constrained by the process-ing power of the array’s storagecontrollers. “Customers scale replication performance in both ourClariion and Symmetrix arrays by distributing data replication acrossa larger number of storage processors,” said Rick Walsworth, directorof product marketing replication solutions at EMC Corp.

With the replication software located on the array, it’s well suited for environments with a large number of servers for several reasons: it’s operating system-agnostic; capable of supporting Windows and

6

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n trendsDATA REPLICATION

These data replication-related trendsare gradually changing data protectionand disaster recovery.

• The use of replication-based data protection is increasing.

• Replication-based data protection ismerging with traditional data protection,and traditional backup products areincreasingly able to manage replicasand snapshots along with backups.

• The integration of replication andapplications to provide application-consistent recovery is on the rise.

• Replication-based data protection isbecoming an important option for pro-tecting virtualized server environments.

• As storage-as-a-service (SaaS) andcloud-based computing become moreprevalent, hosted replication offeringswill become more common.

• Because replication relies on availablebandwidth, wide-area network (WAN)optimization offerings from the likes ofBlue Coat Systems Inc., Cisco SystemsInc., Citrix Systems Inc., F5 NetworksInc., Juniper Networks Inc., PacketeerInc. (now a Blue Coat company),Riverbed Technology Inc. and SilverPeak Systems Inc. are used to comple-ment replication products to preservevaluable WAN bandwidth.

COMPARING DATA REPLICATION METHODS


Unix-based open systems, as well as mainframes (high-end arrays); licensing fees are typically based on the amount of storage rather thanthe number of servers attached; and it doesn’t require any administrativework on attached servers. Because replication is offloaded to storagecontrollers, processing overhead on servers is eliminated, making array-based replication very favorable for mission-critical and high-endtransactional applications.

The biggest disadvantage of array-based replication is its lack of sup-port of heterogeneous storage systems. And unless the array provides a storage virtualization option—as Hitachi Data Systems does for its Universal Storage Platform (USP)—array-based replication usually onlyworks between similar array models. Besides a high degree of vendorlock-in, entry cost for array-based replication is relatively high, and itcould be particularly expensive for companies that have to support a large number of locations. In general, array-based replication worksbest for companies that have standardized on a single storage array vendor.

7

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

Support of heterogeneousenvironments

Performance andscalability

Cost

Complexity

Replication modes

Predominant replication type

Low; only worksbetween similararrays

Depends on thestorage array; verygood in high-endarrays

Requires similararrays; high entrycost; expensive fora large number oflocations

Medium to high

Synchronous andasynchronous

Logical unit number(LUN) or volumeblock-level based

High; storage-agnostic and workswith networked-and direct-attachedstorage

Good; workload is spread acrossservers; limitedscalability becauseof managementchallenges

No hardwarerequired; low entrycost; cost rises pro-portionally to thenumber of servers

Low

Asynchronous

File-system based

High; storage array- and platform-agnostic

Very good

Requires intelligentswitches or inlineappliances; highentry cost; expen-sive for a largenumber of locations

Medium to high

Synchronous andasynchronous

LUN or volumeblock-level based

ARRAY BASED HOST BASED NETWORK BASED

TYPE OF REPLICATION


Almost all vendors of midsized to high-end arrays provide a replicationfeature. The replication products of these leading array vendors have madesignificant inroads and gained market share:

• EMC Symmetrix Remote Data Facility (SRDF) for both synchronousand asynchronous replication, and EMC MirrorView for synchronous andasynchronous replication of Clariion systems.

• Hitachi Data Systems TrueCopy for synchronous replication and Hitachi Data Systems Universal Replicator software for asynchronousreplication.

• HP StorageWorks XP Continuous Access and HP StorageWorks Continuous Access Enterprise Virtual Array (EVA) for both synchronousand asynchronous replication for HP XP and EVA arrays.

8

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

CHOOSING A DATA REPLICATION SOLUTION

1. Selection of a data replication method should start with a business impactanalysis to determine required recovery time objectives (RTOs) and recoverypoint objectives (RPOs).

2. For applications that can’t accept data loss (RTO equals zero), synchronousreplication is required. Heed latency in synchronous replication because it willdrag down application I/O performance. If there’s any risk of latency or unreli-able bandwidth, or for replication beyond certain distances (50 km to 300 km),asynchronous replication is the way to go.

3. Besides the replication mode, application performance can be impacted bythe replication platform. Host-based replication competes with applications forvaluable processor, memory and I/O resources.

4. Have a clear understanding of the bandwidth requirements, impact onbandwidth cost, and how data replication will impact other applications andusers. Clearly understand and take advantage of replication features related to bandwidth such as compression, bandwidth throttling and configurablebandwidth usage depending on the time of day. Consider wide-area network(WAN) optimization devices to preserve bandwidth.

5. Replication products that support heterogeneous environments can sub-stantially reduce cost by supporting less-expensive or legacy arrays. Theyalso limit vendor lock-in.

6. The disadvantage of vendor lock-in of array-based replication is offset bythe advantage of close integration between replication and the storage platformand easier support, eliminating the risk of finger-pointing in multivendorconfigurations.


• IBM SAN Volume Controller (SVC) Metro Mirror for synchronousreplication and SVC Global Mirror for asynchronous replication.

• NetApp SnapMirror for synchronous and asynchronous block-basedreplication, and NetApp SnapVault for file-based replication.

Even though these replication products are similar in many aspects, a close technical analysis reveals subtle differences. For instance, the efficiency of the handshake between primary and target storage sys-tems used during synchronous replication greatly impacts the distancea replication product can support. “Metro Mirror is able to write data tothe target system with a single handshake, enabling it to support dis-tances of up to 300 km,” said Vic Pelz, consulting IT architect at IBM.That distance goes well beyond the 50 km to 200 km cited by otherstorage vendors.

Differences can also be found among asynchronous replication imple-mentations. While EMC buffers data to be replicated in memory, IBMMetro Mirror tracks changes withso-called bitmaps, continuouslytransmitting changes and periodi-cally re-synchronizing the sourceand target to ensure they stay insync. On the other hand, HitachiData Systems uses change journalsstored on disk in its UniversalReplicator software.

“The combination of disk-basedchange journals that are pulled bythe replication targets instead ofpushed by the source makes it extremely resilient, capable of au-tomatically recovering from elongated disruptions,” said ChristopheBertrand, senior director, solutions and product marketing businesscontinuity at Hitachi Data Systems. “Because changes are pulled byreplication target arrays, valuable processing cycles are offloaded fromprimary arrays to secondary target arrays.”

HOST-BASED REPLICATIONIn host-based replication products, the replication software runs onservers so, unlike array- and network-based replication, it doesn’t dependon additional hardware components. That makes host-based replicationthe least-expensive and easiest replication method to deploy.

“Deploying host-based replication only requires installing the replicationsoftware on source and target servers and you’re ready to go,” noted BobRoudebush, director of solutions engineering at Double-Take SoftwareInc. It’s well suited to work in heterogeneous environments, supporting

9

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

“Deploying host-basedreplication only requiresinstalling the software on source and targetservers and you’re ready to go.”

—Bob Roudebush, director of solutions engineering, Double-Take Software Inc.


the widest range of storage options that include both network- and direct-attached storage. While most products support Windows, Linux and Unixsupport is more tenuous and, therefore, platform support is clearly oneof the critical evaluation criteria when selecting a host-based replicationproduct.

On the downside, host-based replication adds processing overhead toservers and the installed replication software carries the risk of introduc-ing unknown behavior. “For critical and high-end application servers, ITmanagers tend to favor array-based replication over host-based replica-tion because it keeps server resources dedicated to the app and doesn’texpose it to potential bugs or flaws in the replication software,” said Lauren Whitehouse, an analyst at Milford, Mass.-based Enterprise StrategyGroup. Furthermore, licensing costs and system administration duties increase proportionally with the number of servers, giving both array-and network-based replication an advantage in environments with a large number of servers. In addition, visibility in host-based replicationis typically limited to source andtarget servers. This is very differentfrom the centralized architecturesof array- and network-based repli-cation offerings that enable a moreholistic view into the replication infrastructure.

The target markets for host-based replication products are typically small- to medium-sizedbusinesses (SMBs) that can’t affordmore expensive replication alterna-tives, enabling them to deploy dataprotection and disaster recoveryarchitectures that, until a fewyears ago, were only seen in larger firms. CA, Double-Take, InMage Sys-tems Inc., Neverfail Ltd. and SteelEye Technology Inc. are some of thevendors that have enabled smaller companies to deploy replication-basedDR and data protection at a fraction of the cost of array- and network-based replication. Although each of these products replicates data from onelocation to another, they differ in features such as efficiency, bandwidththrottling, management, high-availability failover capabilities, platform sup-port and application integration. Only a thorough product evaluation willreveal which product offers the best fit for a given environment.

In addition to these standalone offerings, backup software vendors areintegrating host-based replication into their backup suites with the hope ofexpanding their reach into the lucrative remote- and branch-office data protection business.

10

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

On the downside, host-based replication addsprocessing overhead toservers and the installedreplication software carries the risk of introducing unknownbehavior.


“We see a convergence of DR and data protection, and considerreplication to be a feature and not a standalone product,” said MartyWard, senior director, product marketing for the Data Protection Groupat Symantec Corp. Most backup software vendors already offer host-based data replication options for their backup suites; some examplesinclude BakBone Software Inc.’s NetVault:FASTRecover (formerly Net-Vault:Real-Time Data Protector); CommVault Systems Inc.’s ContinuousData Replicator (CDR); EMC RepliStor to complement EMC NetWorker;and Symantec Backup Exec Continuous Protection Server (CPS) andSymantec NetBackup PureDisk with a deduplication option, as both astandalone product and a NetBackup option.

The main advantage of combining traditional backups and replicationis the ability to manage replicas and backups within a single tool. Asidefrom their host-based replication options, backup software vendors havebeen working on integrating their backup suites with leading storage arrays and network-based replication products to enable customers tomanage all replicas and backupswith the same tool.

“Just like with Continuous DataReplicator, array-based replicas ofsupported arrays are integratedinto the backup application indexand catalog, allowing users to re-store an array-based snapshot bysimply right-clicking it within ourapplication,” said Brian Brockway,vice president of product manage-ment at CommVault. Similarly,Symantec’s Veritas NetBackup is integrated with more than 40 arraysand virtual tape libraries (VTLs), and EMC NetWorker offers tight integrationfor EMC’s RecoverPoint network-based replication product.

NETWORK-BASED REPLICATIONIn network-based replication, the replication occurs in the network between storage arrays and servers. I/Os are split in an inline applianceor in a Fibre Channel (FC) fabric; the I/O splitter looks at the destinationaddress of an incoming write I/O and, if it’s part of a replication volume,forwards a copy of the I/O to the replication target. Network-based repli-cation combines the benefits of array-based and host-based replication.By offloading replication from servers and arrays, it can work across alarge number of server platforms and storage arrays, making it ideal for highly heterogeneous environments. Most network-based replicationproducts also offer storage virtualization as an option or as part of thecore product.

11

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

The main advantage ofcombining traditionalbackups and replication is the ability to managereplicas and backupswithin a single tool.


Contemporary network-based replication offerings are either inlineappliances or fabric based. With inline appliances, all I/Os need to passthrough the replication device. Technically, the appliances terminate allincoming I/Os and initiate new I/Os that are forwarded to the primarystorage targets and, in case of write I/Os, to replicated storage targets. Theinline approach has been plagued by performance and scalability issues.The poster child for inline appliances is IBM’s SVC.

A scalable architecture and plenty of cache have not only enabled it to overcome performance and scalability limitations but, aided by thesimplicity of the inline appliance approach compared to the more com-plex fabric-based implementations, it has become one of the successesin the network-based replication and virtualization market.

In fabric-based replication products, the splitting and forwarding ofI/Os is performed within an FC fabric. By taking advantage of FC switchingand the separating data and control path, it’s the best performing andmost scalable approach. The majority of fabric-based replication prod-ucts run on intelligent switches from Brocade Communications SystemsInc. and Cisco Systems Inc. Eventhough both Brocade and Cisco offer Data Mobility Manager (DMM)for local data center replication,third-party vendors like EMC andFalconStor Software Inc. offermore advanced fabric-based replication products that run on Brocade and Cisco intelligentswitches. A case in point is EMCRecoverPoint, which provides fab-ric-based, asynchronous continuousdata protection (CDP) with app inte-gration that’s on par with commen-surate host-based CDP products. Despite obvious benefits, fabric-basedreplication has seen lackluster adoption.

“Switch-based replication and virtualization have been over-hyped, butthere are people who are working on it and over time it will become morecommon,” said Greg Schulz, founder and senior analyst at Stillwater, Minn.-based StorageIO Group.

LSI Corp.’s StoreAge Storage Virtualization Manager (SVM) straddles theline between inline appliances and fabric-based products that depend onexpensive intelligent switches. The combination of SVM and LSI’s DataPath Module, which plugs into existing Fibre Channel switches to performswitch-based forwarding and eliminates the need for intelligent switches,combines the simplicity of IBM SVC with the performance and scalabilitybenefits of a split-path architecture. HP seems to concur, and is offering

12

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

“Switch-based replicationand virtualization havebeen over-hyped, butthere are people who areworking on it and overtime it will become morecommon.”

—Greg Schulz, founder and senior analyst, StorageIO Group


the LSI product as HP StorageWorks SAN Virtualization Services Platform(SVSP) to complement its host- and array-based replication offerings witha network-based replication and virtualization product.

Even though the market share for array-, host- and network-basedreplication will shift over time, there will be appropriate places for allthree approaches. While each has its own set of advantages and short-comings, specific environments and situations will best determine wherereplication should occur. 2

Jacob Gsoedl is a freelance writer and a corporate director for business systems. He can be reached at [email protected].

13

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE



Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n

FAQ: DR replication

W. Curtis Preston answers some of the most important questions about replication in the disaster

recovery space, including howreplication works, as well as the

pros and cons of synchronous and asynchronous replication.

By W. Curtis PrestonwITH THE INCREASING availability of data replication products,many are turning to this technology as part of a disaster recovery (DR) strategy to address tighter recovery time objectives (RTOs). W. Curtis Preston, an executive editor inTechTarget’s Storage Media Group, as well as an independentbackup expert, discusses replication in the DR space in thisFAQ.

What is replication and how does itwork?Replication is simply the concept of copying data as it’schanging over to another site. So as you change a file that


Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

changes a few blocks, then the idea is that those blocks get sentincrementally to another system that then overwrites the sameblocks on the target system. So you continually have an updatedcopy of what you have on one site living on another site.

How are organizations leveraging replication for DR?For those people who have tried to use tape and traditional methodsof backup and have found that they’re simply not able to meet theirRTOs, those companies have gone to replication for DR, especiallyfor their critical applications. Basically, they’re using synchronous or asynchronous replication so that they constantly have a copy of their data offsite. This way, they don’t have to do a restore whenthey’re in a DR situation. They simply have to switch whichevercopy of the data they’re using.

Can you discuss the pros and cons of synchronous replication vs. asynchronousreplication?It’s not so much pros and cons as it is that each one is appropri-ate for a different application. Synchronous replication is meantto have the updated copy continuously, completely up to date. Itdoesn’t give the acknowledgement of the write to the applicationuntil that write has been copied over to the target site. So that’sgoing to ensure that the target system is always up to date withthe source system.

The disadvantage is that it can create latency in the app becauseif the round trip from the source system to the target systemtakes a long time, you’re going to be slowing down your primaryapplication. That’s what asynchronous replication is for; it goesahead and tells the application that the bite has been stored andthen it asynchronously is copying that over to another site.

The challenge with that is [depending on factors such as howbusy your application is vs. how well your connection is] that youcan be anywhere from seconds to hours behind the source machine.That will result in a loss of data. It’s a tradeoff of how much datayou want to lose vs. how much you want to slow down your primaryapplication.

Generally, people use synchronous only when they’re going local or almost local, where something is only a couple of milesaway. As soon as we start talking about a double-digit number of miles, people generally switch to asynchronous replication because it’s simply too much of a hit on their primary application.

Is asynchronous replication a legit alternative for DR? Can you outline somestrategies for overcoming its limitations?For true DR, you have to be outside of the blast radius. So if thisis Florida and you’re liable to get hit anywhere in Florida with ahurricane, then your DR sites need to be outside of Florida. If thisis California and you want to be outside of the wildfire range,then you have to be outside of California.

As soon as you start doing that, you have to go to asynchronousreplication. The challenge again is “How far behind do you want itto go?” There are a lot of technologies to help make long-termtrips faster to minimize things like the number of round trips. Youdon’t want your asynchronous replication to work exactly like itwas local because when you believe the storage is right next toyou, you do things like ask eight questions when you could ask all eight questions as one transmission.

I would say that it’s absolutely an alternative. What’s the otheralternative: Go to tapes and FedEx them? That certainly isn’t goingto be as good as asynchronous replication. I would add to thatCDP as well, which is basically replication with a back button, so it has the ability to go back in time.

Can you discuss any management bestpractices for ensuring effective replication?At a minimum, this stuff needs to be monitored. Probably thebiggest problem most people have is that they set up that initialreplication and then they’re not doing regular monitoring. I wouldsay that it starts with having consistent replication storage.

There are lots of different ways to replicate: whether it’s storagearray replication, hardware in the SAN replication or software onthe host replication. If at all possible, pick one or at least staywith a common brand. One of the things to do when you select


Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE


that brand is to check out what they have in terms of monitoringand reporting and letting you know when you’re behind, a link hasbeen broken or when the target system that you’re replicating toisn’t there anymore because somebody turned it off.

Also, to make replication affordable, a lot of people are makingthe target system an older system that has been moved out ofthe data center or a less-expensive storage system using FibreChannel in the primary site and SATA in the secondary site. Youmust understand that there will be a significant difference in performance when changing the storage, because it’s where a lot of the performance comes from. 2

W. Curtis Preston is an executive editor in TechTarget’s StorageMedia Group, as well as an independent backup expert.

17

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

0

Data replication


N THE SURFACE, data replication tools represent an efficient strategyfor disaster recovery (DR). A data replication approach relies onferrying updates, transactions and changes across a network—from a primary production server to a secondary backup server.If the primary server goes down, the secondary server can takeover when a failover operation kicks in.

For firms whose recovery time objective (RTO) and recoverypoint objective (RPO) for regaining business-critical functionalityis a matter of minutes, data replication is a truly workable strategy.However, in researching vendors for replication tools, it soon becomes apparent that there are a few catches with the bestand most workable data replication tools:

• The core of any data replication system is the data replicationsoftware, along with the links and hardware necessary for primary

18

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n

toolsWading through replication products is tough

enough, but determining which products best fit your system is even harder. We provide readers with a list of available

replication products and how they differ from each other. By Ed Tittel


Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

and secondary systems to maintain constant contact and com-munication. Some of the major players in this arena include DellInc. EqualLogic (iSCSI SAN arrays remote replication facility), Double-Take Software Inc., EMC Corp. (SRDF/A and MirrorView/A)and NetApp (SnapVault and SnapMirror).

• Where replication data volumes are significant or part ofheavy wide-area network (WAN) traffic, a vital component in anyInternet-based solution is some kind of WAN accelerator. Overtime, the cost of bandwidth for replication becomes a major factor. Consequently, all of the major WAN accelerator players, including Blue Coat Systems Inc., Cisco Systems Inc., JuniperNetworks Inc., Riverbed Technology Inc. and Silver Peak SystemsInc., have extensive partnerships with the leading replicationplayers.

• Likewise, replication touchesdeeply into any enterprise’s stor-age infrastructure, particularlywhere server farms, data centersor cluster/grid computing envi-ronments are involved. That’s why you see storage vendors doing replication business (EMCand Dell EqualLogic), and why somany other data storage vendors(Brocade Communications Systems Inc., Emulex Corp. andSymantec Corp., as well as major system vendors like Dell, Fujitsu,Hewlett-Packard [HP] Co., IBM and Sun Microsystems Inc.) oftencome into this picture as well. It’s possible to get into this marketfor as little as $20,000 to $30,000 (two WAN endpoints with accel-erators, plus replication software). Organizations with more WANendpoints and lots of servers will probably end up spending $8,000to $10,000 per endpoint and approximately $1,000 per server forsoftware (but can drop the endpoint costs for solutions that inte-grate with existing WAN accelerators).

DATA REPLICATION AND WAN ACCELERATORSGiven that the cost of bandwidth can easily consume 30% ormore for annual data replication budgets, making the most of Internet bandwidth between primary and secondary systemslooms very large in the minds of DR and business continuity (BC) planners. This also explains why so many enterprises build

Over time, the cost of bandwidthfor replicationbecomes a majorfactor.


data replication solutions around combinations that work withtheir existing WAN and server infrastructures, which togetherdictate what kinds of replication solutions get purchased and deployed.

Sometimes, when always-on or hot recovery sites are con-tracted from third-party suppliers, those site operators will specifyreplication solutions. Because duplicating IT assets essentiallymeans multiplying IT costs by the number of recovery sites setup, this gets expensive fast. This also explains why it’s crucial to limit recurring costs for any replication solution, where band-width and personnel costs quickly swamp equipment and software outlays. As you examine your options for datareplication, the most importantconsiderations that emerge will most often come from asystems integration perspective.After that, you’ll also want tocompare features and functions,talk to reference accounts andmake sure your candidates’products will work well in yourproduction environments, oftenthrough a pilot test of somekind.

Very few companies havethe luxury of crafting suchproducts from scratch, andmust work with or around ex-isting components and tech-nologies. That’s what makes replication so interesting for thosebusinesses and organizations that decide they need it for disas-ter recovery or BC reasons. It’s not a difficult option to select, asmuch as it can be a challenging option to implement. That’s whyworking from the WAN accelerator out to the servers and storageelements appears to be the most commonly practiced and productiveapproach.

Take Silver Peak Systems as an example of this approach. Itsdata replication products embrace a range of offerings, includingDell EqualLogic, Double-Take, EMC, NetApp and Symantec. Thus,companies with existing investments in those tools and tech-

20

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

As you examineyour options fordata replication, the most importantconsiderations thatemerge will mostoften come from asystems integrationperspective.


nologies will find it easy to add (or build upon) Silver Peak WANaccelerators as part of a coherent data replication strategy. Similarinvestigations of offerings from Blue Coat, Cisco, Juniper Networksand Riverbed will turn up similar alliances and concomitant pos-sibilities for implementation. That said, it’s probably a good ideato carefully manage your costs for software and integration toprevent them from ballooning out of control.

Ultimately, the best data replication solution is one that inte-grates most easily (and inexpensively) with existing infrastructureinvestments, and that shows itself to support BC/DR planning recovery objectives. Because the WAN accelerator componentsare so important to controlling recurring costs for data replication,they’ll often drive other elements that come into this picture.That’s why we recommend that you start with your WAN accelerator vendor when it comes to scoping out a suitable product, or that you base your choice of WAN accelerator at least in part on the data replication options it supports. 2

Ed Tittel is a long-time freelance writer and trainer who specializes intopics related to networking, information security and markup lan-guages. He writes for numerous TechTarget websites, and recentlyfinished the 4th edition of The CISSP Study Guide for Sybex/Wiley.

21

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

aStorage Essential Guide to Replication for DR22

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n

T THE HIGHEST LEVEL, there are three main types ofdata replication commonly used: application-based,host-based and storage array-based data replication.We could get even more granular and further subdi-vide array-based replication depending on whether ittakes place at the controller level, the storage-areanetwork (SAN) level or is controlled by an appliance.

Purists could argue that SAN and appliance-basedreplication aren’t “true” array-based replication be-cause they’re independent of the disk array, but forthe purpose of this article, we can agree that replica-tion takes place at the storage level rather than beinghost or application based. What distinguishes storage-level replication is that it relieves the application andserver resources from the processing overhead associated with replication.

Disk array-baseddata replication:

PROS AND CONSThere are three main types of data replication:application based, host based and array based.

Learn the differences between the threetechnologies, how to get started with array-based replication and find out

about available products.

By Pierre Dorion


THE DOWNSIDES OF ARRAY-BASED REPLICATIONA few years ago, it was a lot easier to outline the downsides of array-based replication; it was a very low-level technology that replicatedblocks of data without much ceremony. Many times you had to takeapplications down to preserve data integrity because the applicationwasn’t aware of the replication process. The technology offered very little support for heterogeneous storage hardware, which made itpricey. It typically worked at the array controller level, and remote replication was costly due to net-work bandwidth requirements andoften required some proprietaryform of protocol conversion. Mir-roring or RAID 1 is probably one ofthe earliest and best known formsof controller-based (array-based)replication. It’s also known for itscost and limitations.

Not that these days are longgone, but data replication optionsare now much more flexible and affordable. Once reserved for enterprise-class organizations, there are now many offerings available to small-and medium-sized businesses (SMBs) without breaking the bank. Thishas made array-based replication a very popular option because it’scentralized and operating system-agnostic. However, interoperabilitybetween most vendor solutions still remains a challenge.

That being said, if a very low-cost replication solution for a limitedamount of data is what’s needed, array-based replication is often stilltoo expensive for small environments or small remote offices.

GETTING STARTED WITH DISK ARRAY-BASED DATA REPLICATIONThe best way to get started with array-based replication is by first answering a number of high-level questions:

• What are you trying to achieve? The need to implement array-basedreplication may be to address slow backups, increase the frequency ofbackups, capture frequently changing data, reduce or eliminate tapemanagement, etc. If host-based or application-based replication is tooresource intensive, array-based replication is likely the right approach.If large volumes of data are replicated, host-based or application-basedreplication can also interfere with network traffic.

• What do you need to replicate? The answer to that question can influencethe cost of the solution so it’s always good to address this realistically. Be-cause technology is deployed to support the business (in most cases),there should be a business requirement for replication benefits ratherthan a “nice-to-have” capability.

23

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

Mirroring or RAID 1 is probably one of the earliest and best knownforms of controller-based(array-based) replication.


This takes us back to a very familiar discussion about data availability,recovery time objectives (RTOs) and recovery point objectives (RPOs),which ultimately drive the need for a particular technology. There mightbe a need for a certain application to have access to the latest possiblecopy of data. In such cases, a technology like EMC Corp.’s RecoverPointis well suited. The I/O is split at the SAN level and written simultaneously ontwo different supported storage arrays, local or remote. This is donetransparently and provides continuous data protection (CDP).

However, the ability to access point-in-time copies might be desirable.Data corruption or unintentional data deletion are often cited as situa-tions where this functionality is required. Snapshot technology is a goodfit to create multiple point-in-time copies without requiring an amountof disk space that’s a multiple of the space the original data occupies byonly replicating changed blocks. NetApp Inc.’s Snap software suite offersa comprehensive set of options, including local/remote snapshots, mir-roring, vaulting and application-specific snapshot capabilities.

APPLICATION-AWARE REPLICATIONApplication-aware replication is probably the most significant developmentin the field of array-based replication. The integration of replication withapplications has made it possible to create copies of the data while theapplication is up and user access is maintained. Whether the productleverages snapshots, mirroring or volume-based replication, the abilityto create a replica of a particular data set without affecting user accessis very much aligned with today’s availability requirements. This capabilityis also far superior to the traditional daily backup because it can supportmuch tighter recovery point objectives by allowing multiple daily copiesor even continuous protection.

LOCAL VS. REMOTE REPLICATIONOne more significant feature of array-based replication is the ability toprovide local or remote copies of the data. While this capability certainlyisn’t unique to array-based replication, it’s not as intrusive or resourceintensive as host- or application-based replication, which provides anopportunity to replicate more frequently.

VENDOR CHOICES IN ARRAY-BASED REPLICATIONThere are many vendors with array-based replication offerings and theircosts vary. Some of the best-known vendors include:

• EMC’s TimeFinder, SRDF, MirrorView, SAN Copy and SnapView• IBM’s Tivoli Storage Manager for Advanced Copy Services• Hitachi Data Systems Universal Replicator, TrueCopy and

ShadowImage • Hewlett-Packard (HP) Co.’s StorageWorks suite of software

24

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE


Some vendors with specialized offerings have smaller market sharesbut offer innovative data-protection technologies with an array-basedreplication component. For example, Data Domain Inc. (an EMC company),leverages data reduction through data deduplication as a backup datastorage target enhanced with array-based replication for disaster recovery (DR) purposes.

That said, with the exception of EMC’s SAN Copy, which replicates atthe logical unit number (LUN) level, and Hitachi Data Systems’ UniversalReplicator, which leverages TagmaStore, heterogeneous support remainsthe main challenge for array-based replication.

The line between host-based or application-based replication and array-based replication isn’t as clear as it once was. The emergence of virtualization appliances is becoming the answer to interoperabilitybetween heterogeneous platforms while remaining off the host. 2

Pierre Dorion is the data center practice director and a senior consultantwith Long View Systems Inc. in Phoenix, specializing in the areas of businesscontinuity and disaster recovery planning services, and corporate data protection.

25

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

uStorage Essential Guide to Replication for DR

SUALLY, when a customer asks a reseller to provide a product for datareplication to a disaster recovery (DR) site, the reseller thinks, “Well,that’s good news and bad news.” The good news is that the customerwill need extra storage. The bad news is that the extra storage maybreak the budget, and the customer may have to put the purchasingprocess on hold to find additional funds.

Beyond potential budget problems, there’s the technical difficulty.Setting up an efficient and easy-to-use replication topology is tediousand time-consuming. Managing and implementing replication via prod-ucts such as EMC Corp.’s SRDF and Hitachi Data Systems’ TrueCopy andUniversal Replicator is a very complex undertaking; it’s also costly, anddifficult to adopt without some sort of professional services engagement.

26

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

n

ChannelPerspective:

Benefits of thin replicationReplication coupled with thin provisioningis making its way into mainstream storageproducts. Learn how thin provisioning can

enhance your replication product. By George Crump


Despite both of these problems, a replicated copy of data for DR is nolonger a “nice to have” for most data centers, it’s a requirement. So howcan a solution provider address the need for replication, have a satisfiedcustomer and live to tell the tale?

Until recently, there was no easy answer. But next-generation repli-cation techniques that will make life easier for resellers and their cus-tomers are on the way. Companies like Compellent Technologies Inc.and 3PAR Inc. are leveraging thin provisioning capabilities into theirreplication offerings. Of these two companies, Compellent has gone thefurthest by providing advanced quality-of-service (QoS) and bandwidthshaping. Armed with these “thin replication” products, you’ll no longerhave to spend day upon day at your customer’s site configuring thereplication topologies with complex Fiber Channel over IP (FCIP) inter-connection schemes, full disk-based copies, little bandwidth control and cumbersome management interfaces.

WHAT MAKES THIN REPLICATION THIN?Traditional storage-area network (SAN) replication involves a significantplanning cycle to determine which protocol to use, acquire the neededconversion hardware and decide how much bandwidth is required. Andthe actual implementation is complicated because those scripts have tobe written, tested and fine-tuned. Finally, ongoing management is a chal-lenge, as no data center is static; changes to the application infrastruc-ture will mean changes to the replication environment and a call to thestorage reseller for support.

One of the benefits of a thin replication product is that it provides theability during the initial replication phase to replicate only the data thathas been written to a given area rather than the entire allocated amountof capacity. Most other techniques require a block-for-block copy. Forcontinuous replication, only the changed data sets are replicated to thealternate locations. This gives your customer the ability to conservecostly bandwidth segments that would normally be overtaken by repli-cating needless amounts of allocated but unused capacity (white space).The target systems can be deployed with a diminished capacity, savingcustomers money, and allowing you to sell only the bandwidth and storageyour customers need at the time they request it.

Thin replication reduces the replication process to just a few mouseclicks. It’s no longer necessary to undertake large scripting efforts to automate replication; instead, the process is driven through simple andeasy-to-create replication templates. You can install the product withoutneeding office space while you make it work.

QUALITY OF SERVICEMost replication techniques don’t let you scale bandwidth prior to in-stallation, though it can be accomplished via integration or added soft-

27

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE


ware functions. Another benefit of thin replication is the variety ofunique tools that allow you to model the customer’s bandwidth andprovide a QoS level—capabilities missing from other replication tech-niques. Being able to look at your customer’s source data and modelthe required replication bandwidth up front, before they procure thebandwidth and/or target systems, means you can sell only what yourcustomer needs for their DR replication.

In addition to enabling users to estimate the amount of bandwidth adata center will need in the future, thin replication provides the ability tothrottle bandwidth usage throughout the day and monitor that bandwidthusage over time. This gives your customer the flexibility to replicate datawithout burdening business operations during peak hours.

Even if your customers don’t start with replication on the initial installa-tion of these products, these tools let them continuously model their datasets over time. When they’re ready to implement replication, they canprovide you with all of the necessary data to do so. In a typical scenario, you would need to use host-based tools along with array-monitoringprocesses to determine the upfront and ongoing change rates. The QoSoffered in thin replication lets storage administrators change a numberof very useful parameters, such as:

• Scaling replication bandwidth up and down during peak and off-peakhours to ensure adequate usage of the links based on customer service-level agreements (SLAs).

• Giving priority to one data set over another based on the business-critical nature of the data contained within a given volume. This allowsan administrator to provide, for instance, a higher spec to an SAP volumethan a file server one.

On the topic of QoS, legacy replication offerings have been slow toadapt to handling any real QoS concerns solely by the functionality con-tained within the replication and array areas. The task of collecting themetrics and then performing automated functions based on those met-rics would be very slow and painful, based on the number of areas youwould need to go to gather information (such as servers, storage, SAN, IPnetwork, wide-area network and so on). This makes traditional replicationimplementations that have to meet a certain performance or utilizationrequirement very risky for resellers to guarantee.

THE BENEFITS OF INTERCONNECT PATHWAYSUsing simple networking technologies to provide the interconnect path-ways between the target and remote systems allows for even more flexi-bility in the replication solution. With thin replication, there’s no need forcomplex FCIP engines on both sections of the replication segments. Thedeployment is as simple as plugging into the iSCSI connection on thestorage system and then into your customer’s network. It’s then a matter

28

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE


of using already available pathways to gain access to the target systems.While FCIP is supported even by thin replication products, most customersgenerally default to iSCSI as the transport of choice because it’s cheap andeasy to use.

In more legacy-like deployments, you would need to purchase the FCIPengine and integrate it into your Ethernet and Fibre Channel networks. Fora robust and redundant deployment, these devices are typically doubledto ensure physical-layer high availability in the replication link. The FCIPdevices are fairly complex to configure and maintain over time.

THIN REPLICATION IMPROVES VALIDATIONThin replication also enables storage administrators to test and verifythat the DR and business continuity sites are in solid working order with-out affecting the production volumes at the source location. This allowsa customer to perform a simple procedure to determine that the recov-ery time objective (RTO) and recovery point objective (RPO) can be metwith no impact to the primary facility. These verification processes couldbe added to the daily, weekly or monthly checks to determine data in-tegrity at the remote location, delivering a simulated DR process.

In a traditional replication environment, this functionality is almostnonexistent. Testing DR plans in real-time requires a significant time investment from a number of IT groups within your customer’s organiza-tion or the purchase of specific aftermarket products that provide thatfunctionality.

With thin replication, no matter what topology the customer selects,the replication process can be configured, managed and monitored fromone management console, minimizing the number of full-time employeesneeded to maintain and manage the system.

Disaster recovery is no longer a “someday” project, as most medium-sized businesses require some form of near-real-time data replication.The good news is that the benefits of technologies like thin replicationhave significantly lessened both the cost of entry and the technicallearning curve. The advanced features and functionality built into thinreplication provide you and your customers with a flexible, simple andcost-effective replication product to meet many enterprise needs. 2

George Crump is president and founder of Storage Switzerland, an IT analystfirm focused on the storage and virtualization segments. Prior to foundingStorage Switzerland, Crump was chief technology officer at one of the nation’s largest storage integrators, where he was in charge of technologytesting, integration and product selection.

29

Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE


Dat

a re

plic

atio

n to

ols

DR

rep

licat

ion

FA

QR

eplic

atio

n

alte

rnat

ives

Arr

ay-b

ased

rep

lica-

tio

n: P

ros

and

cons

Ben

efit

s o

f th

in r

eplic

atio

nSTORAGE

Check out the following resources from our sponsor:

A Practical Guide to Disaster Recovery Planning: The Basics to Getting Started

Real-Time Data Protection for Microsoft Hyper-V

Next Generation Backup for Remote Offices

http://www.doubletake.com/uk/Pages/default.aspx

http://www.bitpipe.com/detail/RES/1256322565_19.html



Documents

ESSENTIAL GUIDE TO Replication for · d Storage Essential Guide to Replication for DR ATA REPLICATION as a means of data protection has seen contin- uous and increasing adoption since