With natural and man-made disasters, it is quite a challenge for a business to continue functioning.
With the growing dependency on IT to support business operations, the risk of losing data has become critical and hence importance of recovery post disaster remains a top priority for organizations. Vamsicharan Mudiam, Country Manager, Cloud Solutions, IBM, India/South Asia speaks to Dataquest on how to save critical data during disaster via private or public cloud-based, on-site or off-site data backup services
What according to you is a disaster?
Recovery time objective (RTO) and recovery point objective (RPO) effectively define the business impact of a disruption in service and broadly define disaster recovery. A disaster is any disruption in service, not necessarily something that happens when a disaster is declared by the country. It can be a break that arises due to an unforseen situation. About 50% of disasters occur because of weather changes - not necessarily an extreme one like a hurricane or fire, but inaccessibility due to cables cutting down because of strong winds.
In the past IT used to be more inside the enterprise to better your function but not necessarily visible to your end customer. At that time, 'disaster' was an extreme situation and the IT people or your business people required help to access information. Today, most enterprises interact with customers and have discussions on IT. So any service interruption is actually a disaster today. There are people who lose thousands of dollars for every minute of breakage of service, whether you are collecting data, collecting bills or you are selling retail merchandise.
A dedicated disaster recovery or shared disaster recovery (DR)?
If you take cloud away from the equation, a dedicated DR is a hot site. It is more expensive and often, you assume you'll never use DR just like you don't take an insurance policy to encash it. You are not creating a DR to actually use it but as a fall back.
When you go for a dedicated DR, you invest in bandwidth, hardware, replication of data etc. It's an expensive proposition if one needs DR only for one or two weeks of usage. Hence, people following a shared DR model. Infrastructure is shared and replication is done on certain servers for different customers. The advantage is cost but the time consumption varies. The biggest problem is replicating your data, it takes about 24 to 72 hours, i.e. the time taken in a shared infrastructure to get your business back up and running. It is not speedy enough but it is cost effective.
How can virtualisation help in DR?
Virtualization is basically adding that new flavor of ability to save money and a crucial amount of time to get your system up and running. It is almost like the best of the two worlds. Virtualized Server Recovery (VSR) introduces cloud behavior into shared infrastructure. It creates an ability to automate the deployment much faster, like the classical cloud. It enables you to automate and progress much faster. It is not as expensive but is almost as good as a dedicated set up. That is where this whole angle of disaster recovery on the cloud is coming up. Firstly, we are leveraging the existing model of shared infrastructure investment for creating disaster recovery setup. Secondly, we are leveraging the advantage of cloud where we are able to vitualize infrastructure and create these instances on demand, so that the best of these two combinations is creating this disaster recovery on the cloud.
As server recovery solutions, we have both physical and virtual servers which we recover because there are some customers who are running work better off left on physical due to licensing norms or until they migrate completely.
You don't work with a customer's estimate. Please elaborate.
We have a network estimate rule which we give to customers to estimate their network, required bandwidth etc. The ability to understand how much of an application is written and not written is very dependent on changes that occur on the application, new initiatives etc.
Suppose you acquire an average of a thousand customers in a month and you developed a bandwidth for your CRM application and DR. But when you roll out a free offer for the next three months, you might on-board 1500 customers. We are not saying we won't help, but we want the customer to own that because the ability to understand their application is best left to a customer. That's the reason we give broad range of RTO and RPO of minutes up to 4 hours.
Are Indian enterprises aware of disaster recovery solutions and using it?
The fraction of people using DR is huge but I don't know the exact number of customers we have. The uptake is huge on virtual server recovery itself, and so, people are aware of these solution. We also organise our own programs where we conduct programs focused on this topic.