The problems are common across organizations:
· Reduce costs
· Simplify data movement
· Improve RTO/RPO objectives
· Simplify DR
Over and over, different company, same story, and the reason why the top technologies in 2010 will be the innovative solutions which enable companies to reduce costs while improving service levels. This is why cloud computing is starting to make an impact, since it lets companies outsource some areas of operation to reduce costs.
The enabling technologies which enable cloud companies to build out their service offerings for data protection are Virtualization, Continuous Data Protection, and Dedupe. Storage virtualization removes the physical boundries of data placement and location, and creates the abstraction from physical resources. Continuous Data Protection changes the physics and the rules for backup and recovery, and dedupe makes the storage, replication and archiving of data assets more efficient. Cloud service providers are leveraging these technologies and using them to help companies offset costs. Let’s make up some new acronyms and call these services BAAS (backup as a service) and DRAAS (disaster recovery as a service).
BAAS and DRAAS will actually be combined into a new service called CDPAAS, which is my acronym for continuous data protection as a service, which lets companies outsource the entire backup and DR functions.
Even if these services are still provided internally, organizations need to ensure these innovations are in their data protection plans, or they will simply be left behind. The cost containment they provide can create a competitive advantage for many companies. If MY IT is more efficient than YOUR IT, then I can provide more cost effective products and services than the YOU can.
Let’s look at two companies side by side and compare cost structures:
|Company A||Company B|
|DR||Array based Replication and tape||Fabric based dedupe and replication|
|Server RTO SLA||24-48 hours||15 min|
|Server RPO SLA||24 hours||15 min|
|Time to DR||1 week||1 hour|
|Cost of downtime||$1000 hour||$1000 hour|
Now let’s use a recent disaster news story and put the production location in Haiti, and the DR site in the Dominican Republic.
Both Company A and B lost many key people, and their entire DR plan is still under the rubble. What are their prospects for survival as they rebuild? Company A’s primary data is safe at the DR site, but most of the data is still on tape. Only their critical data is available on the other storage array, and since their key people and recovery plan are lost, no-one at the DR site knows in what order or at what consistency point their inter-related applications need to be at for recovery. Also, due to the flood of calls for resources at the recovery provider, no hardware is currently available to rebuild their application servers. Even if the hardware was available, their contract calls for a 48 hour SLA per server (48 x 1000=$48,000.00 per server), and a 1 week disaster recovery time (7 x 24 x 1000 = $168,000.00), and there is no-one around that knows how to bring up the applications anyway.
Now let’s look at company B, who lost a similar amount of IT resources and people in the disaster. Since their data was continually replicated and recoverable to virtual servers at the cloud service providers site, the service provider was able to bring up their applications within the contract window to known good consistent snapshot recovery points. The provider took an hour to recover all applications virtually as they failed over to the known good points in time.
Total costs are meaningless in this example for company A, since company A will probably out of business as company B takes over their customers.
Which company would you rather depend on for your job?