I have decided to start a new series called “Questions To Ponder” which I will push out on some Fridays. The idea is to take a controversial topic and ask a few relevant questions so that CloudAve readers can ponder about it on their weekend. The first in this series is about the role of integrated stacks as we move towards more and more services world. There are big hardware vendors like IBM, Oracle, Cisco, etc. who are pushing the idea of tightly integrated stacks as the ultimate solution for all of IT’s pain points. Their argument is that procuring different hardware and software components from various vendors lead to integration hell resulting in higher costs and lack of agility. Instead they push their ultra-expensive integrated stacks as the alternative solution. They argue that the seamless experience offered by the integrated stacks from a single vendor is worth all the additional costs. They have their own ROI studies to prove their points.
On the other hand, cloud computing is pushing us towards a loosely coupled services world where these integrated stacks have very little relevance because the end users just consume the services and leave the integration headaches to the service provider. Service providers use a multi-tenant model for efficiency and use a scale out approach to meet their scaling needs. Integrated Stacks have very little relevancy for these service providers providing mostly modern web centric services. They also have the necessary expertise to handle all the operational hassles.
However, organizations are not going to throw away their legacy applications and move to modern web services overnight. The move is going to be gradual and we will see a more hybrid infrastructure during the intervening period. Integrated stack vendors understand this shift towards the services world and have repurposed their marketing messages (of course with some tweaks to their offerings) to argue that integrated stacks help organizations meet the challenges of the hybrid world. However, if any organization wanting to embrace the loosely coupled services future will run into the lock-in risk if they embrace the integrated stack for their hybrid needs. It is not easy to move to a loosely coupled services world if the existing infrastructure is on top of integrated stacks offered by a single vendor unless of course they want to move to the services offered by the same vendor. I seriously doubt any traditional vendor will completely disrupt themselves to offer services that match the ones offered by the new disruptors in the market.
Under such a scenario, is there any place for integrated stacks as we move more and more towards the loosely coupled services future? I do agree that integrated stacks play a critical role in some niche scenarios but does it mean anything otherwise? Should we buy into the marketing campaigns of the integrated stack vendors or stay clear of them to prepare for the future? In other words, does the idea of integrated stacks have any relevance any more? These are some of the questions to ponder.
PS: It is easy to nitpick on my arguments above but what I expect is answers to bigger philosophical questions than nitpicking based on niche scenarios
Disclosure: IBM is a client