For at least one year integrated infrastructure stacks have been sprouting, big vendors are pushing them strongly and customer are looking in depth at this “new” proposition.
What is a stack and Why
An integrated infrastructure stack is comprised by:
- computational power (often in a blade form factor);
- networking stuff (sometimes converged networking for both storage and network);
- an enabling hypervisor (VMware is the best choice at the moment but Hyper-V and Citrix are growing);
- a single management tool (only a dream at the moment);
- support and services;
The professed goal of integrated stacks is the promise of a better overall TCO of your IT infrastructure thanks to:
- a single vendor to deal with for consultancy and support;
- a better integration of all the components of the stack because they are fully engineered to work together from the beginning;
The real goal the vendors have is to sell more to extend the footprint inside the customers’ datacenters. Some vendors try to improve their sales by simplifying the proposition. Indeed, the biggest threat of integrated stacks is the lock-in with one vendor and the risks associated with it.
The are a lot of stacks about from many vendors. Some of these stacks aren’t credible (Oracle is an example), some of them are just to-be (see HDS) but some of them are here and they are trying to gain some traction in sales (VCE-acadia, Dell, HP). And where is IBM? I think that IBM doesn’t like the stack concept itself because they prefer to sell you a service to configure a stack than a preconfigured stack (probably because they are the truly inventors of stacks!)
Where the stack wins
The stack makes sense when:
- the customers doesn’t crave for the best of breed in his infrastructure but to something that just works,
- he has not enough IT staff resources and he looks for something simple to manage,
- he would like to cut the set of vendors.
This kind of customer expect to lower management costs and to avoi integration issues when chooosing an integrated stack. It’s not a a bad choice even more if your IT infrastructure is static and has a predicatble life: without an explosive data and computational growth and while having standardized processes and applications to maintain.
Probably, the TCO model of these customers hasn’t the hottest spot in the infrastructure related issues.
Where the stack fails
The stack isn’t a solution when the customer looks for the best performance, the best features, the maximum control and when he has enough resources, quality as well as quantity, to manage them. The most growing datacenters (i.e.: cloude services providers) aren’t paying attention to integrated stacks, instead they are looking for non locking, best, cost saving technologies!
TCO model for this kind of customers are more infrastructure related.
Why some non-stack-integratd storage vendors are growing faster?
Stacks are not winning right now!
Many customers don’t like to be locked and a lot of them are still looking for the best solutions, not for those that are just simple to buy! If you compare sales from all the vendors you will find that:
- Acadia customers are growing but not so fast! and CISCO blade platform has more installations with NetApp then EMC!
- Compellent, 3Par, NetApp are growing more than the market: they are all not-integrated-in-stacks storage vendors!!!
- Virtualized environments need fully virtualized storage architectures to work at their best and primary stack vendors aren’t ready yet.
Three clues don’t make an evidence but they may help you when considering integrated stacks!