From Wikipedia, the free encyclopedia
“Converged infrastructure packages multiple information technology (IT) components into a single, optimized computing solution. Components of a converged infrastructure solution include servers, data storage devices, networking equipment and software for IT infrastructure management, automation and orchestration.”
I am not a converged infrastructure hater. I agree with the value proposition that converged infrastructure brings. VBlock from VCE, FlexPod from the reference architecture of NetApp & Cisco, HP with Cloud Matrix. The afore mentioned converged infrastructure solutions are marketing machines. The big players in this space have strong relationships with their current customer base and are using converged infrastructure as a one stop shopping menu for their gear.
Where I work today converged infrastructure makes perfect sense. We have boxes and boxes of equipment bought for projects and then no people to implement the hardware. The operational tasks of maintaining the equipment are not sexy and a lot of people don’t seek out to run firmware upgrades. The expected lower operating costs by letting a company like VCE per test everything is very appealing. The other players could do the same but they just don’t wrap it up in a nice SKU.
The other major reason why converged infrastructure is getting to be so popular is around risk. Companies want to put the risk back on the vendors. If something goes wrong, one phone number and one throat to choke is the mantra. Does this really happen? It seems like it’s Dilbert cartoon waiting to happen to me but I get it. All of the converged infrastructure companies mentioned at the beginning of the article are market leaders, have a proven brand and deep pockets to help current customers if something where to go bad. There is not a lot reasons to feel bad about picking one of these solutions around risk but finger pointing always seem to the enviable.
The issue that I have with converged infrastructure is that it’s solving a problem that these companies helped to create. Traditional SAN architecture using separate server and storage tiers and is complex to scale and manage. To fix the problem they offer rip & and replace and lots of consulting money to help maintain this new beast of the data centre. Why not look at it a different way? This is way I am a big fan of Nutanix, small modular 2U building blocks with scale-out performance. Relative a new company on the scene but has gotten great press at VMworld 2011 and at Tech Field Day 8. Nutanix has solved the problem that traditional vendors are fighting and using the same mindset that made Google the company they are today, they got rid of the SAN! Using local storage in their “nodes” they use the “Nutanix Distributed File System” that localizes data to boost performance and reduce storage network traffic. It’s a dense solution with 48 cores, up to 768GB RAM, and 21TB of both flash and spindle disks. The numbers will only go up when Nutanix releases their Gen 2 products this fall.
With Nutanix I like the fact that I don’t need to go into deep planning for something that is three years down the road. If I need more storage or compute I can buy another 2U node and I am off to the races. It allows for projects and budgets to grow as they need, just add a node for linear scalability. With vSphere already integrated into their stack I have some concerns with that but I wouldn’t virtualize with anything else at the present time so I can live with that requirement. I am not sure about all the available tools around their solution but I what I have seen so far looks good. It looks like they actually had a usability person design the user interface.
I would love to see Nutanix where I work but we have strong existing relationships with most of the current companies mentioned here. Those companies today are really are extension of our IT department. It will take a lot to unseat the incumbents in the data centre but I think Nutanix has a compelling story.