Is the Future in the Clouds?
So what’s the scoop on cloud computing? Are our data centers going to disappear into the cloud?
I’d suggest that these are the wrong questions.
First, as in any industry, there are the usual leaders. IBM, Google, Microsoft, VMWare, and others seem to have gotten out of the blocks fast, but so far none of these companies seems to have a distinct advantage over the others. In addition, the little guys continue to have a great growth market. For instance, one LinkedIn post asked whether Spring Source’s recent acquisition of Hyperic would enable it to challenge giants Microsoft and IBM.
Lately discussions about implementing clouds have focused on chargebacks. Can corporate enterprises charge individual departments for use of internal clouds? For that matter, have cloud vendors developed a winning revenue model? We seem to have returned to the same old argument of how to pay for IT.
Nik Simpson summed this up nicely in a May 5th post:
“IT infrastructure purchases in enterprises are frequently tied to business unit budgets that fund them. As long this infrastructure was physical, tying budgets to asset via asset management systems or CMDBs was still possible.”
With the cloud’s dynamicity and pooled use of resources how are enterprises going to manage cost and budget allocation to specific business units and processes? Are monitoring and metering systems being launched that address this issue? Is this holding up cloud adoption in large enterprises that need a level of maturity in chargeback systems before they bring the cloud in or scale out?” For many organizations, lack of chargeback is a symptom of a wider problem; the inability to accurately model costs for IT. If you can’t say how much a particular application costs, it’s hard to create a usable charge back model.”
Security is another issue. How will corporate, private, and regulated data be protected while it is on an Internet cloud? Creating your own corporate cloud would seem to answer these basic issues, yet the frequency of hacks into many traditional corporate networks leads security experts to raise flags about the security of private clouds. It seems that we have yet to understand the ramifications of cloud operations, and the tools to monitor and manage a cloud are only now being developed.
Cloud computing applications are still in their infancy, and like any group of infants, some applications will mature faster then others. Right now traditional HPC, scientific, and database applications are rapidly tapping the performance of cloud computing for high-speed (both in CPU and I/O) highly parallelized semantic analytical applications.
Managed storage systems and IO devices that can’t match the processing or telecom speeds lag behind other processes moving to clouds. Further, most applications are yet to be tuned or converted to run in the cloud so there are still many years of work for mainstream IT organizations to envision the benefits of cloud computing for their organizations.
Future cloud users could include the military for real-time battlefield analysis. Artificial intelligence, robotics, weather forecasting, deep-sea exploration, and, of course, the entire gaming industry make good cloud applications. It’s ironic since some of processing techniques used in these fields made possible the processor clustering designs essential to cloud computing. Who knows what other contributions could come from the clouds? Captain Kirk’s universal language translator or finding a solution for global warming?
So what does cloud computing mean for our data centers?
First, Moore’s Law will continue and as such will demand ever more creative ways to house the new technology as its form factor changes.
Second, data centers will continue to reach the end of their useful life in 10 years (give or take a year or two).
Lastly, our industry will continue to need more professionals who know how to provide high availability, reliability, resiliency, modular ability, portability, scalability, and much more.