Site Recovery Manager 5.0 ups the ante

One of the best things to come out of VMworld 2011 was the announcement of the latest version of VMware’s vCenter SRM (Site Recovery Manager). While previous iterations of SRM have brought little more than bug fixes and relatively minor feature additions, SRM 5.0 packs in a number of critically important features.

Those features, including built-in host-based replication and automated failback support, substantially increase the capabilities of SRM — even in small to midsize business environments where it had previously gained little interest due to its relatively high cost of implementation. The introduction of replication to SRM 5.0 is also a continuing reminder of VMware’s transformation into a de facto storage company as it tacks more and more storage array functionality into its product line.

SRM before 5.0

At its heart, Site Recovery Manager was and still is a site failover automation engine. Prior to SRM 5.0, there was nothing that SRM did for you that you couldn’t do by yourself manually. However, that’s certainly not to say that it was without any value — far from it. By automating the site failover process, you could shave hours or even days off of the time that it would take to reconstitute a virtual infrastructure at a warm site.

SRM also made it extremely easy to perform thorough tests of the failover process at any time without disrupting production services. That testing capability is arguably SRM’s strongest feature. Without an automation engine to run a site failover, it’s often so time consuming to test that all but the most diligent organizations will rarely make the time to do so regularly, thereby opening themselves to the real possibility that their failover system will fail when they most need it.

However, earlier versions of Site Recovery Manager did have some serious drawbacks. The most critical was a requirement that the underlying storage of the site SRM was protecting had to be replicated to the recovery site using ABR (array-based replication). ABR can be terrifically expensive to implement — often prohibitively so for smaller organizations.

First off, it means you need to have two compatible SAN infrastructures, which immediately limits the number of organizations that can afford it. Many SAN vendors also require you to apply expensive licensing to their hardware to enable the replication features. Beyond that, there’s the substantial intersite WAN bandwidth required to allow high-volume replication to function — often dwarfing the cost of the second SAN over a few years.

Finally, SRM licensing isn’t cheap. Though it was initially sold as a CPU socket license applied to each host at the production site, VMware has since transitioned it to a per-VM license that runs in the neighborhood of $320 for each virtual machine when long-term support is included. That’s a lot to ask for a tool that also requires you to implement an expensive SAN-based replication infrastructure.

Larger enterprises that were already utilizing SAN replication as part of their business continuity solution saw SRM’s comprehensive automation capabilities as an easy way to deliver on testing requirements without requiring a huge amount of manpower. However, many smaller organizations balked at the combined cost of buying SRM and implementing the ABR that it required.

What’s hot

VMware vCenter Site Recovery Manager 5.0 brings a raft of new features to the table. However, the one feature that’s rightfully catching everyone’s attention is the ability to leverage HBR (host-based replication) either by itself or in combination with ABR, thereby making the purchase of a second SAN infrastructure entirely optional.

This feature is a transformative one for Site Recovery Manager. If you’re trying to offer business continuity to a small business with 20 VMs and would like to replicate them all to a single vSphere host with a large amount of internal storage, that’s now feasible without requiring the use of third-party software like Veeam or vReplicator. That’s not to say that those third-party products are suddenly obsolete — in most cases they offer backup-related features that SRM doesn’t. However, the failover automation capabilities of SRM that initially drew in large enterprises are just as attractive to smaller businesses without the staffing or skill set to monitor and manage multiple systems. SRM helps out there by drawing all of the management of production and disaster recovery functionalities within the same pane of glass.

This new HBR capability also has implications for large enterprise — even ones who already use ABR for their primary data centers. Multisite enterprises with small virtualized infrastructures distributed at remote offices can leverage vSphere Replication to replicate edge-site VMs back to centralized data centers, either as a business continuity functionality or to make the process of migrating those servers back to the head end easier.

On top of all of this, VMware has actively been courting VMware-based cloud services providers to use SRM’s HBR capability in new DRaaS (disaster recovery as a service) offerings. Already, companies like FusionStorm, Hosting.com, Iland, and Veristor are offering or preparing to offer such cloud-based solutions. This opens the possibility for true business continuity to organizations without the capital or real estate to implement a warm site.

The replication functionality isn’t all that’s new, either. Other great features include the ability to define a migration workflow to handle planned site migrations, automated reprotection after failover, automated failback, and more granular VM dependency definitions in site failover plans (ensuring that machines with interdependencies start in the correct order).

However, some of these features (site failback, for example) are currently only compatible with ABR configurations and won’t yet work with the new vSphere Replication feature. Similarly, HBR configurations are soft-limited to 500 VMs per SRM infrastructure (versus ABR’s 1,000 VMs). Like many other scalability limits in VMware products, you can expect these to be raised in future releases of SRM as customer demand dictates.

Putting it all together

New functionality offered by vCenter Site Recovery Manager 5.0 dramatically widens its applicability beyond large enterprise. As a result, I predict it will become much more popular than previous versions. However, SRM 5.0 isn’t just interesting because of the capabilities it brings to bear. It’s also a sign of a continuing trend on VMware’s side of slowly transforming itself into a storage company. When you combine SRM 5.0 with the recently released vSphere Storage Appliance, VMware can now offer many features that would previously have required two stacks of SAN hardware. Granted, there are presently limitations in both of these products that would make the inappropriate for large enterprise, but this is absolutely a sign of things to come.

This article, “Site Recovery Manager 5.0 ups the ante,” originally appeared at InfoWorld.com. Read more of Matt Prigge’s Information Overload blog and follow the latest developments in storage at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.