Two Ways To Think About Business Continuity

On 19 March 2013 by Pete Petersen
Business Continuity

Image source: http://siliconangle.com

One of the foremost concerns of businesses of all kinds of business continuity. If the financial and other industries, engaging in business continuity planning (BCP) is a dreaded task. BCP, in this case also includes Disaster Recovery (DR) planning (or DRP, in some circles). It’s expensive, time consuming, and there is rarely immediate financial benefit from doing so–in fact, the opposite. In most organizations, the key personnel that you would task for such an important effort are already time-taxed to overflowing.

There are ways to make the entire BCP effort much simpler to accomplish. VDI is one of them.

A True Disaster

Consider this.

When managed desktops and applications are centralized, they have the ability to move. For example. Let’s pretend that my data center in Miami goes to lunch because of hurricane, flood, fire, terrorist attack, or whatever. Pretend that the entire data center is now gone. If I were planning my VDI (and other compontent) infrastructure correctly, the users will need to merely log back into their broker site and re-launch their desktop and applications again. Any unsaved data is lost, of course, just like it is when your local machine crashes.

But let’s get real, our local machines crash significantly more often than we lose a data center.

Think BCP/DRP

 That being said, here are the items that need to be in place for the users to experience little/no downtime in the event of a true disaster.

  • User data dispersal and/or replication (consider Cleversafe with a CIFS front-end as a dispersal example and Microsoft DFS as a replication example)
  • Application data dispersal or replication (SQL mirroring or log shipping as examples)
  • VDI farm redundancy (identical farms in more than one place)
  • Load balancing between farms (NetScaler DNS-based Global Service Load Balancing (GSLB) as an example–in this way, there aren’t as many idle resources)
  • Redundancy within farms (think N+1, or 25% overage, as examples)
    • File data
    • Database
    • VDI compute and farm data
    • Directory Services
    • Storage
    • Virtual machines as needed
    • Configuration data
  • Networking components (firewall, switches, application filters, etc.)

Note that whenever implementing a new technology, or refreshing an old one, BCP, including DRP, should at all times be forefront of our thoughts.

In a traditional scenario, if a data center is lost, then users (or user devices) will have to figure out where all their data is again. That’s a lot of variability existing out at the end point.

In a centralized (VDI) scenario, in a well-implemented VDI environment, the user needs merely to log in again to another location where the settings already point them to the right location.

Let’s Talk Small

So that works well in a large environment. Having come from a large, regulated environment that serves a quarter of a million users, this kind of thinking becomes almost intuitive. It’s talked about regularly. There are required planning forms to keep current. And regulatory agencies that regularly want an audit of BCP readiness.

So think of a small or medium environment. Let’s pick 500 users as a starting point. We’ll get to the small environments in a minute.

In a 500-user environment, options are more limited. Organizations normally won’t be able to afford to spin up several data centers of their own in order to satisfy the needed BCP requirements. But they could co-locate. In fact, this may be a better strategy anyway.

If your organization is still putting servers in closets and under desks and letting Production users run there, it may be time to think about pulling those out and relocating them to a safer environment. There are many accredited data centers that would be happy to host your servers for a nominal fee. Many do full-rack, half-rack, or even quarter-rack deals that can satisfy organizations of almost any size.

Once you’re into a data center and have that figured out, now it’s time to move to the second data center. This is where your BCP planning will tap your best minds. All of that data, the systems, user experience, and so on, will need to be reproduced in the alternate site. And if you’re smart about it, both sites are load balanced so that in the event of a disaster, users are effectively none the wiser and can continue working as normal–BCP testing is happening every day.

Now, think of an 80-user environment, or even a 30-user environment. What can possibly make sense at that scale? BCP is just as important to the 30-user environment for their business as it is for the 250,000-user environment for their business. Business must continue in order for the organization to survive.

Conclusion

Virtualization technologies, including server, desktop, and application virtualization, make BCP planning much more flexible, cost-effective, and manageable.

If you’d like to explore the possibilities further, please contact us for a discussion.

In addition, these two excellent articles will add some insight.