Some organizations with more than one location choose remote data replication for disaster recovery. There are a lot of factors to consider when deciding if this approach is right for your needs. David Chapa, senior analyst with the Enterprise Strategy Group, outlines the pros and cons, what you need in place to facilitate this approach, how virtualization fits in, and the major vendors in this space. Read his answers below or download the MP3 of the conversation.
Table of contents:
>> What do you need to consider first before deploying remote data replication?
>> What technology do you need?
>> What are the biggest challenges with using remote data replication for DR?
>> How does virtualization fit in?
>> What are the major vendors in this space?
So, if your organization has two sites, and you want to replicate data at one to the other and vice versa for disaster recovery, what’s the first thing to consider? In other words, are there some scenarios where this just won’t work?
I think it really depends on the type and size of the organization. When you look at enterprise vs. a small- to medium-sized business (SMB), I think that question can be answered differently. Also, in some cases, just because a company is an SMB, they wouldn’t require a high-bandwidth solution.
It is important to consider bandwidth requirements, the type of files being replicated, what kind of WAN optimization solutions are available, and what kind of replication technology they will be using.
Can you discuss what you’ll need to have in place to set something like this up? Obviously, you’ll need to have storage at each site and the ability to perform replication, but what about beyond that? For example, is some sort of WAN optimization necessary?
Is it always going to be necessary? Probably not always. Is it nice to have so you can maximize the efficiency of your WAN pipe? Definitely. When you are talking about disaster recovery, it really comes down to what’s important. You need to take a look at your data, understand the importance of your data, and determine what data you need to recover in a certain amount of time to recover your business.
If you look at say, a video editing house where they are working with very large files that don’t compress very well—they might not be a good candidate for optimization, but it is certainly something many organizations should consider and test to determine if they will get a benefit for their investment.
The biggest challenge is determining what the most important data is for your company. If you can afford to replicate everything: great. I think a lot of companies replicate everything to avoid the daunting task of classifying their data. Especially for SMBs and smaller companies, data classification can be a real challenge. I’ve done it as a consultant, and it is not fun.
However, once it is done it allows you to plan a lot more effectively, for DR but also how they protect their data on-site.
We have some research that shows that a lot of companies perform agent-based backup of virtual machines. But recently, we’ve been seeing a lot of technologies from the ISVs and VMware that allow you to back up from outside the virtual machine.
If you have a large environment of virtual machines, you have to start looking at the best ways to protect that. Is it array-based replication? Can I replicate using the disk that is hosting the virtual machine in a way that is recoverable at the destination?
One other thing to consider is Changed Block Tracking [which VMware introduced in vSphere 4] that allows users to replicate only changed blocked.
It comes down to importance of data, recovery time, and bottom line, it comes down to budget.
I’ll probably miss something, but some examples include EMC Corp., NetApp Inc., Hitachi Data Systems Corp., and Hewlett Packard Co. All of those companies have a way to set up a disaster recovery for the virtual infrastructure. The other thing to look at is what happens if you don’t have a 100% virtual environment? What happens if you are 25% virtual and 75% physical? How do you manage that?
When you start thinking about that stuff, you’ll be looking at some of the solutions that allow for physical-to-virtual recovery. I’ve done a lot of research as of late, including a backup-as-a-service market landscape that we published a couple months ago, in which I looked at all of the backup services and the majority of them offer physical to virtual, virtual-to-virtual, or virtual-to-physical type of recovery. That opens up a lot of options for users to manage data across tiers.