In my last piece, we discussed what cloud computing meant exactly and tried to unravel the lingo behind the technology. In effect, cloud computing did sound like a technological panacea from utopia for all problems for elastic computing and storage.
However, it is not so. There are some significant technological and political challenges that need to be overcome, before riding the wave on cloud.
This is probably the most significant challenge that an enterprise has to overcome before trusting their sensitive data to a third party storage provider. Cloud service providers have to establish to their customers that the data is secure. This is done by providing a ‘Service Level Agreement’ (SLA). Security does not just mean ‘away from the prying eyes of hackers’. It also means that the data is physically secure. That is, all the data of the enterprise is not lost because of a disk failure or fire accident. Clouds typically handle by caching data and backing up across data centers.
Geo-Location of Data
This is more of a concern for government agencies on the political boundaries where the data is stored. Some governments have restrictions that its data can be stored only within its political boundaries to avoid sensitive data, falling in the hands of its political enemy (or WikiLeaks :P). In this case, geo-redundancy, which is one of the actual advantages of moving to a cloud-based infrastructure, is a counter-point for its adoption!
TheSLAundertaken by cloud providers typically indicate that minimum availability of the data that has been uploaded to the cloud. Though 100% availability is usually desirable, typical rates are in the range of 99.9% to 99.95%. Some also have penalty clauses that would reduce the cost of cloud subscription if the provider does not meet the terms offered in theSLA. But the typical penalty (to the provider) is in terms of more cloud compute/ storage. So, the possibility of becoming a ‘cloud millionaire’ using these penalty clauses is usually remote.
Not all cloud technology is plug-and-play. Even enterprises that want to walk the ‘cloud way’ may not be able to do that because of incompatibility or legacy technologies that the enterprises may be wired (locked) in. For example, probabilities of finding a dancing penguin in the middle ofSaharadesert is slightly higher than finding a cloud infra provider for running COBOL apps that take care of data in mainframes.
I really don’t want to be the bad guy here. I spoke of advantages of cloud in the previous installment and I have been only complaining this time. I hope no one started an immediate migration to cloud based on my previous article.
Cloud technology is new and the logic is sound – “don’t pay for the resources that you don’t use, if you can”. Cloud offers a way to do this, with other benefits. An enterprise has to carefully weigh the pros and cons, and existing political scenario before choosing a cloud provider, if at all you need one.
Despite all the seemingly big disadvantages, I am reminded of the ‘network is the computer’ slogan of erstwhile Sun Microsystems. There were slightly ahead of their time (at least, in their slogan). These are times when storage and computing are actually on the network!
My forecasting skills are not known to be legendary. But we can safely conclude that the cloud is here to stay, for a while. Jump into the bandwagon, only if you want to!
This is a guest column by Sriram V Iyer, Co-founder, United Mobile Apps (check out their story by clicking here). To read his earlier column on Nokia’s plans with Qt, a cross-platform toolkit, click here. Also, check out his piece on the iOS v/s Android battle with respect to the problem of fragmentation faced by app developers, by clicking here.
Also, do share with us your thoughts & views on this story by writing to us at firstname.lastname@example.org.