Just A Thought About A Cloud Internet

Cloud computing comes to NERSC

Cloud computing comes to NERSC (Photo credit: Lawrence Berkeley National Laboratory)

On Monday, I flipped through Flipboard and couldn’t understand why halfway through Flipboard threw a tantrum.  The tantrum went on a long time, and so I had to refrain from using Flipboard until late in the night, adhering to eastern standard time.  Today, the New York Times’ “Amazon Cloud Service Goes Down and Takes Popular Sites With It” piece confirmed Flipboard’s Monday tantrum.  Of course, not only Flipboard but other huge, gigantic kind of services that use Amazon for cloud services had also experienced the Monday case.  Monday, Amazon acknowledged that cloud servers experienced degraded performance and problems with the Northern Virginia data center(according to the New York Times).

I’m making a big deal of this incident is to point out that cloud services (as in singular) isn’t invulnerable to outage.  In fact, I believe (but not knowing for sure) that cloud services might have a tendency to breakdown even more since the complication of keeping everything tidying within a cloud is not that simple.  Of course, when a cloud is doing everything right, it can be way more resilient than non-cloud services.

Amazon is one of the biggest cloud players in the market.  Whenever Amazon has cloud problems, it shows that cloud services can be just as vulnerable as non-cloud services.  I think cloud services (as in singular) is more resilient in many ways (e.g., data redundancy, scalable computing, etc…), but whenever it sneezes million of services will catch a cold too.  This prompts me to ask a question, how much bigger a cloud needs to be so that a common cold won’t be powerful enough to infect a cloud?

I imagine, a company Y which provides cloud services to have a backchannel which connects to a cloud infrastructure of company X, and when company Y experiences a cloud cold, company X’s cloud infrastructure would kick in to cure the cold for the cloud of company Y.  The billing will reflect the rescue operation, therefore there is an incentive to setup this sort of backup cloud infrastructure between these gigantic cloud companies.  Imagine this scenario would be interconnected as large as if there is a separate Internet, but this Internet is responsible for cloud services only.

Of course, just like the Internet, if a company isn’t too careful in applying security protections and so forth, a cloud Internet can be just as infectious as the regular Internet.  A cold in a cloud Internet has the ability to infect way many more services than a regular cloud.  If this is the case, why would I think a cloud Internet is better?  Resiliency is the answer if a service does care about not ever going down for a cloud outage.  I think the cloud Internet would definitely be large enough to absorb most cloud outages and allow time for the cloud outages to be fixed and reenergized.  Then again, I might not know what I’m talking about since I’ve never actually operated and engineered a cloud.  (Almost forget to bring this up, perhaps encryption would be the panacea for data security since cloud backchannels might get rather even more murkier in term of who is in charge of whose data.)

Source:  http://bits.blogs.nytimes.com/2012/10/22/amazon-cloud-service-goes-down-and-takes-some-popular-web-sites-with-it/?partner=rss&emc=rss

Advertisements

One thought on “Just A Thought About A Cloud Internet

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.