If you think that private clouds are just doing public cloud-like things within the data center, you're dead wrong. As the hype builds up around private clouds, the approaches to building these yet-to-be-defined virtualized systems are really left up to who's building them. There is no one approach, nor is there a killer technology in this space as of yet.
Let's take a look at a few major reasons why we'll build private clouds. First, we love to control things and we can't directly control the existing public cloud providers. Second, we may have some special security and legal issues that prevents us from placing our information outside of our firewall. Finally, we can't sell all our existing hardware and software on Craig's List, so we might as well figure out something to do with it.However, if you think that building a private cloud is just a matter of Google searching for private cloud technology providers, you have a rude awakening coming. While the number of private cloud and virtualization startups increase every month, there is still no one single killer technology out there that will make moving to a private cloud both cool and easy.
That was the conclusion made by Dmitry Sotnikov in his recent post "No real private clouds yet?"
"In a nutshell, private clouds are Amazon-like cost-effective and scalable infrastructures but run by companies themselves within their firewalls... [However], there are companies working on elements of private cloud systems of the future but there is nothing close to being a real solution with Amazon-like acceptance."
While a number of upstarts in this space defy the recession, we've yet to find that one technological path to travel, and that one killer solution. I suspect that it's going to be a while before one emerges, albeit some are promising.
In my journeys as a consultant, I'm finding that each enterprise has its own set of private cloud requirements, and they look very different than public clouds. Thus, most solutions in 2009 may be a bit of SOA, with some virtualization and identity management tossed in, but not an out-of-the-box solution.
Another aspect is the value of private clouds. While many are making claims that the use of private clouds will drive better resource utilization, the metrics in this arena are still a bit cloudy when considering the investment required. Moreover, the larger data center vendors, including IBM, Cisco, Sun, and Oracle, have their eye on the emerging private cloud market and are currently creating solutions and bundles to dive into this market while the hype is hot. However, many fear that the rush in building this technology could create many proprietary solutions, and thus vendor lock-in. Again, if the private clouds don't have some direct and measurable value, there is no reason to build them.
While cloud computing is hot, private clouds could be where the money is spent in 2009 and 2010. It will be interesting to see how all of this evolves. Watch this space for my take.If you think that private clouds are just doing public cloud-like things within the data center, you're dead wrong. As the hype builds around private clouds, the approaches to building these yet-to-be-defined virtualized systems are really left up to who's building them. There is no one approach, nor is there a killer technology in this space as of yet.