Connect with the world's leading business experts.
Get instant access to their expertise via world–class Q&A, Research, and Events.
0
In a virtualized cloud network, what are the challenges of establishing DR (disaster recovery)?
Events
- Social Media and Content Marketing For Business Q&A Feb 14 @ 11 am PT
- #TNLive Radio: Workforce Marketing & Recruitment Feb 14 @ 4 pm PT
- The Rise of Pinterest in B2B Feb 15 @ 11 am PT
- ERP – Priming Your Business to Deliver Value From Strategy to Operations Feb 15 @ 1 pm PT
- How Not to Coach Your Salespeople Feb 16 @ 1 pm PT







5 Answers
Actually in a properly configured virtualized cloud disaster recovery is pretty simple. If anything happens to the virtual server, simply "reboot" or re-image the virtual server and the system is back up in minutes vs. hours or days.
Hi Erica
What exactly do you mean by "virtualized cloud network". Can you give us an example?
Thanks
John
By virtualising the production servers, clusterring physical host servers and backing up virtual server disk files (vhd or vmdk) will make DR Job as simple as copying files. However the challenges will become how detail of the recovery level would be. For example, file level, mail box level, object level in AD or even bit level recovery or maybe just server level of restoring the whole vhd or vmdk files.
Also virtual disk snapshoots sometimes would create issues if it is not managed properly, some recovery jobs would be involed such as merging snapshoots. Personally, I don't recommend to perform disk snapshoot on production mechine.
It very much depends on the ability of storage environment (if it is cloud, is it stored in the same region as your cloud servers) and the possible tiering and latency involved on restoring a server.
When virtualizing for the cloud, you should take into account what your recovery point objective (RPO) for data loss will be and the service recovery; recovery time objective (RTO).
Your RTO may be less than your RPO if you have infrastructure to provision virtuals immediately or you split your workload across sites. This is a challenge for many.
Today's capabilities by HP/IBM/EMC and many others provide synchronous and asynchronus data replication across frames at the speed your network can handle. Second challenge, depending on the amount of replication for various data sources, ensure you have adequate bandwidth.
Another challenge is commnication/network links redundancy and ensure your points are terminated by different carriers.
When you talk about DR, it is obviously different then High Availability (HA). You can have HA and if virtualized, you can recover the virtuals quickly to meet your RTO. However in a disaster of your active site to your DR site, you need to look at all the levels involved to restore business to its original state. Technology is the enabler but you need to look at user functionality testing; user acceptance testing; system testing and others to validate the data now available to your DR site is what your end-users/customer need to complete transactions and continue to generate revenue for your company.
Thanks.
Gonzarelli
Answer This Question