This year's model
Virtualisation is starting to make an impact in the disaster recovery space. Every disaster recovery project undertaken by mid-market service provider Brennan IT in the past six months has been predominantly based on virtualisation, says managing director David Stevens.
"Our entry-level solution is that we replace physical servers with virtual servers in one of our data centres. Combined with that is either live data replication or very close to live," Stevens says.
"For a minimal monthly outlay they get access to a whole bunch of virtual servers, and then the pricing changes if they need to enact the disaster recovery plan. It's a much better use of money and resources than buying a lot of extra hardware and having it sitting idle in a data centre somewhere. A lot of our clients have been so impressed with how the virtualisation side of things works that they've switched their core business apps over to virtualisation as well."
When it comes to disaster recovery, the size of an organisation isn't as important as how critical its data is.
"It's really about how quickly a customer needs to be able to access that data. Financial records are of particular importance and compliance regulations are still really driving a lot of the disaster recovery projects out there," Stevens says.
"If you're in a financial services role then disaster recovery becomes important even if you're a one- or two-person company."
Brennan IT's cloud computing reseller program lowers the barriers to entry for channel players looking to offer enterprise-grade disaster recovery, says Nick McMenemy, Brennan IT's head of marketing and product management.
"We have a range of commercial models, from the basic referral to a full white label "provider-in-a-box" solution. There's been a lot of interest from the traditional reseller community in branding their own flavour of service, but it's taking time to filter through because there's obviously a degree of awareness and learning required," McMenemy says.
"It's a very interesting proposition for the channel when you consider the alternative of doing it yourself from scratch - the capex outlay, the rackspace and connectivity, the management overheads and the need for new skills.
"Then they've got to provide a reasonable service-level agreement and a helpdesk. Or they come to us and we do it all for them, on a pay-per-use basis. That's a model that translates very readily into the channel."
Virtualisation is revolutionising the disaster recovery market for the channel by enabling new services that are affordable for smaller organisations, says Leonard Chan, future technologies manager with Macquarie Hosting.
The cost savings and flexibility associated with virtualisation are allowing SMBs to switch from the traditional "primary and backup" disaster recovery model to a "dual production" model. This allows customers to achieve a higher level of availability.
"A lot of customers already use ‘active/active' virtualisation in their production environment. It's not often we come across a pure physical server play these days. Typically a customer may have all their active servers running in one facility and when anything happens things are transferred across to the secondary facility," Chan says.
"The trend is now towards dual production servers, so they continually distribute the load between servers and data centres so you don't put all your eggs in one basket. This is part of a shift from disaster recovery being a reactive implementation to making it more proactive. And the key to this is virtualisation. The fact is that when you use technologies like virtualisation it really makes disaster recovery much simpler and easier to enact."
The biggest hindrance to the dual production model is not the hardware or the services but the high-speed data link required between two active servers in two separate data centres.
"Fat pipes can be cost prohibitive for SMBs, but we can tweak the interaction between dual production servers to reduce bandwidth requirements," Chan says.
"Obviously this is yet another area where the National Broadband Network is going to level the playing field, so customers even at the small end of town can still afford all the services the big players enjoy, but without the capital expense. The NBN will mean big things for SMBs and the disaster recovery space."
Along with bandwidth, storage is a key component when it comes to disaster recovery systems. Just as with servers, the storage space is also reaping the benefits of virtualisation, says Scott Morris, director of partner sales with storage vendor NetApp.
"Right now I'd say business mobility and agility are pushing virtualisation more than disaster recovery," says Morris.
"At the moment people seem to be more interested in how to get more from less but, when you sit down and look at it, improved disaster recovery is one of the true benefits of a virtualised environment.
"You can spin up virtual servers from backups, and virtual storage is no different. As fast as you can bring up servers we can virtualise storage - we call it thin provisioning and it can be a real boost to a disaster recovery plan.
"Virtualisation can be a risk because you're combining separate physical servers into one system, generally with network-based storage such as a NAS or SAN device," says Morris.
"This means the impact of a single failure can be greater, unless you put some other protection mechanisms in place. That's really where we're seeing a massive groundswell of opportunity, as all of those resilience factors that we've built into our platforms tie very strongly into virtualisation and disaster recovery."
Racked with opportunity
Advanced disaster recovery systems which involve synchronous data replication between sites obviously rely on fat pipes, but the load can be reduced by falling back to semi-synchronous data replication - running every hour or every day rather than every keystroke. Site-level high availability technologies can also significantly improve basic disaster recovery plans.
"There are obvious benefits to handling disaster recovery in the cloud, but I don't think many customers appreciate site-level technologies such as snapshots. This lets you easily return a server to the last known good state without having to resort to a tape or online backup," Morris says.
"Another site-level technology of interest is RAID. Employing RAID technology involves spreading data across multiple hard drives to protect against drive failure, but many customers don't appreciate the limitations of common configurations such as RAID 5.
"The weakness of RAID 5 is that it only protects you against one hard drive failure, then you need to swap out that drive and rebuild the RAID. Rebuilding the RAID could take up to 12 hours if you're using 500GB drives. Meanwhile you're at risk of data loss should a second drive fail. Alternatives such as RAID Dual Parity offer an extra level of protection."
The flexibility of virtualisation-based disaster recovery processes is enabling a new breed of dynamic disaster recovery services, says Viktor Babkov, managing director of Business Continuity Asia Pacific (BCAP).
BCAP is the sole Australian distributor of Double-Take software, which is designed to offer real-time backup and automatic failover capabilities for physical and virtual servers regardless of the hardware or software.
"Virtualisation is a great enabler of disaster recovery, allowing rapid scalability of needed resources on demand," Babkov says.
"We can move existing infrastructures into the cloud, we can set up non-intrusive mirrors on applications. We can marry an existing data centre infrastructure with existing infrastructure on the premises.
"There's so much that can be done, but SMBs in particular often don't appreciate this flexibility. This is where the channel can step in to educate them about the potential of these new technologies and thus open up new markets."
-1.jpg&w=100&c=1&s=0)
.png&w=100&c=1&s=0)
