File this blog under “serendipity”.
At the time of this writing, I’m sure we have all heard
about the data disaster concerning the Sidekick cellphone. The Sidekick may
have the distinction of being the first cloud cellphone, and while never a
threat to the BlackBerry or iPhone, had a loyal customer base and a solid
following in the under-25 market, the barometer of hip.
All user data was
stored by T-Mobile (working with Microsoft subsidiary Danger) in the cloud, it does not sync with a PC - an interesting feature of the Sidekick as well as other smart phones such as the Palm Pre. Last week, Sidekick users started suffering from
outages, and then seeing their contact information disappear. Passwords,
contact lists, pictures, you name it – gone. Microsoft is working to see if
the deleted data can be restored, and after a few tense moments earlier this week, it now looks promising.
When defining the cloud (and believe me, every day on my
Twitter account at least two people ask that question) there are as many
definitions as there are people to describe it. I believe
the common feature is that the primary software and storage is stored remotely
from the users, and is effectively black boxed from those who are using it. We
don’t care that there are 1,000 servers or 3,000 servers – or (for the most
part) where they are. As a cloud
customer, I care only that my storage capability is available when needed. And
for goodness sake, make sure that the data is secure and if it is damaged for
some reason, well, I shouldn’t know the difference.
The bigger question here for the future of services tied to cloud computing is disaster recovery (DR). It would seem to me a natural use of the cloud if there ever was one. Catastrophic loss of data due to an act of God
(think earthquake), or man (think September 11) takes out your server farm - your customers still
need their data. I’ve worked all over the US and it's pretty clear that 9/11 did a lot for the DR evangelists among us. The Sidekick incident might be the rallying cry for the mobile ecosystem.
In California, companies I’ve worked
for have been planning disaster recovery since probably the day after the 1906
Earthquake. Several CIOs have told me that it wasn’t a major priority on
the East Coast, but 9/11 changed all of that forever. The need to have your
backup data stored in a physically removed location from your production data
is crucial, and the cloud is perfect for that. The advantages are clear: no/low equipment and maintenance
costs, a potentially vendor agnostic hardware setup, and the ability to conduct
full testing performance or load testing without impacting your production
environment.
However, a recent Symantec survey of 1000 IT managers and came
up with some stunning facts, including this one: 35% - yes, over one in three - of an organization's servers
are not included in a DR plan. No performance or disaster recovery testing.
The news for the systems that are tested isn’t going to
prompt parades. In the same survey, 30% of the respondents
reported failure in their testing, whether it be incomplete backups or slow
performance. This led Internet News to proclaim “DR testing is a
mess”.
I could be on an island on this one, but think that performance testing should be considered an aspect of DR (or vice-versa). Imagine the scenario that your servers are disrupted by a 6.0 quake - known in California as a minor shake up (smile) - I would expect my DR planning to demand fast on-boarding of the backups and the performance of the cloud servers should be (ideally) as fast as the primary in-house servers.
Risk management has always been a concern for cloud advocates and others see this as the hype bubble being pricked. But as in many cases in life, trouble can be opportunity in work clothes. Time will tell, but I don't see this as a case to be made against using the cloud for data storage. Rather this is a good reason why a solid, risk-managed cloud strategy for rapid, high-performance disaster recovery spells can spell opportunity for businesses and enhance consumer confidence.
More on this the next post.

Now this is hghly recommeded post for me. I will surely email this to my friend.
Regards
Natalie
Posted by: Raleigh Home Remodeling | November 09, 2009 at 06:21 AM
Wonderful post... The loss of data dampened the believe everyone had in the clouds... Only time will tell!
Posted by: Acai Berry Pure | January 19, 2010 at 12:08 AM
I could be on an island on this one, but think that performance testing should be considered an aspect of DR (or vice-versa). Imagine the scenario that your servers are disrupted by a 6.0 quake - known in California as a minor shake up (smile) - I would expect my DR planning to demand fast on-boarding of the backups and the performance of the cloud servers should be (ideally) as fast as the primary in-house servers.
Posted by: generic viagra | March 11, 2010 at 01:42 PM
Excellent post.Generally I do not post on blogs, but I would like to say that this post really forced me to do so.
Posted by: aion kinah | March 11, 2010 at 11:47 PM