In response to a few thought provoking questions from a colleague on whether Oracle VM, VMware or IBM would be better prositioned to lead virtualization of Java, I had to form a few responses and decided to share them with everyone, and gather insights and comments from others who read my blog. So after much rumination on hot technologies (all biases aside as best as I could) I can share what’s been stewing up in my mind for the past few months.
So, running WebLogic on the hypervisor is compelling, but I doubt many companies will want to migrate to Oracle VM in order to obtain this advantage. Check out the recent Gartner report that VMware is alone in the Leader Magic Quadrant for virtualization, so this is no slam dunk for Oracle, Microsoft, or any other vendor. But challenges are ahead for Oracle in virtualization on hypervisor, as one article puts it, “Either they (Oracle) promote VMware, and abandon their own product, or they abandon their customers, but keep their product.” I haven’t really expressed much of an opinion here, as much as I have doubts about customers reaction to the technology that’s available. I’m not so much of a virtualization guy as I am an IdM guy, but time will tell, and with any luck Oracle may relax their position of resistance against virtualization from their Palo Alto brethren.
But going beyond the datacenter, now many customers have the option to run Java apps in the cloud rather than their own infrastructure using VMForce. My bets are that history will repeat itself, and this trend will only continue as companies abandon in-house server farms and infrastructure, and as Nicholas Carr aptly describes in The Big Switch: Rewiring the World, from Edison to Google, will opt for deployment to Cloud machines much the same way companies abandoned generating their own power and began using centralized electricity provided by the power grid in the early part of the last century.
With that said, IdM technology is in for a roller coaster ride as the tidal waves of change come and we look at how to manage and scale IAM services across a broad spectrum from internal IT to private clouds to the public cloud for partners, customers and employees. It is looking like the cost and complexity of extending Federated SSO across multiple protocols (not all customers will have SAML, WS-Sec) will be a hassle unless you factor in the potential of cloud services and a hub & spoke model. It makes me wonder if IdM will go the way of the centralized power grid and Cloud Services (IaaS, PaaS, et al) or maybe it’s already happening. And as Coby Royer points out in a recent blog post, “I can install the old style IAM tools, this is missing a huge opportunity for cost savings—putting standard infrastructure for IAM into the “drinking water” is the wave of the future.” In an economy like this, that logic is becoming much easier to buy into than say in 2008 before the recession started to hit IT budgets.
As an old hand at Oracle IdM (going on 10 years now) it is a bit hard for me to digest, but my instinct tells me that survival means adapting to the seas of change rather than trying to run from them. There is a bright future and a lot of pent up demand in cloud services, where new models will soon overshadow the shortcomings of client/server and internet architectures. The old school IAM stacks are not going away anytime soon, but the IdM professional will need to learn new models and standards to keep pace with where this industry might be heading.
Anyways, time will tell. Leave a comment if you think differently.