A few weeks ago, @JimPeluso sparked a great question, "How will a Vblock change IT in 5 years?"
Let's examine how everything is done currently. An IT Manager is given a virtualization/VDI project with a budget of $900k. The IT Manager gathers his network, storage, and server teams and discusses what can be done. Or perhaps he grabs his one lonely systems admin and asks what can be done. No matter what size the company, this is going to be a long battle with a bunch of different vendors. Each person does their part and narrows down the playing field to 3 specific vendors they want running in their data center. Each vendor has a different plug, gimmick, and sales pitch, but in the end, all an IT Manager cares about is getting the job done and helping the business succeed. The IT Manager listens to the high level pitches then hands it off to engineers to develop a solution. Every vendor, both storage and servers, are going to ask the same questions, "What kind of workload are you running?", "Do you know the IOPS?", "Do you know what kind of performance you need?" and so forth. Yet, the same questions are being answered multiple times to several different companies. A week or so flies by and the IT Manager receives multiple solution responses and discusses it with the engineers over the next few weeks to decide we want Vendor X for storage and Vendor Y for servers.
Now it's time to fulfill that PO, and it takes another 2-3 weeks for all the equipment to arrive at the data center floor. Now the engineers are given the task of racking, stacking, cabling all the equipment together. We're all the correct sizes of cables ordered? Is time a factor? So do the engineers quickly cable it all together to get it functioning or do they take their time to do proper cable management? All the equipment is hooked up, time to hit the power button and make sure everything is talking. I don't know about you, but in my past experience, it's a rare occasion when everything works on the first try. After another week of burn-in time, the engineers can start doing a proof of concept (POC) on the new equipment. They begin taking the low hanging fruit or test/dev network and start running VMs. After another month or so (depending on IT lifecycle) more production VMs can be put on the new equipment. At this point, it would have been nice to have all parties of storage and server talking on the same call to verify best practices with different types of hardware, technologies and protocols. If Tier 1 apps begin to under perform, where does the troubleshooting begin and how is it alleviated? Who's throat are we going to choke? The IT Manager just spent $900k in new equipment, where does his/her return on investment (ROI) come in to play? At this point, it's a fairly good guess that it's taken 6 months to almost 1 year to get a good amount of production servers on the new equipment.
This is a big problem for Service Providers wanting to do multi-tenant environments and need the ability to scale quickly. Service Provider X just landed a huge account that wants to throw 500 VMs on a hybrid cloud. How quickly can the CTO get his team to procure new equipment and tie it in with current infrastructure?
A Vblock changes the game at this point. Instead of gathering teams of engineers to piece-meal a solution and answering the same answers
over and over again, why not buy a product that encompasses a full stack? After VCE is engaged into a deal, a vArchitect or team of vArchitects will listen to your needs and create a solution in a rack (or few racks). The Vblock is guaranteed to deliver what it promises. Whether it's IOPS, a specific amount of VMs, or a particular need such as SAP/VDI. This new evolution of procuring IT equipment is made easier and with greater ROI. The IT Manager can spend $900k on a Vblock and start seeing ROI in as little as 3-4 weeks compared to 6 months to 1 year. This complete integrated stack relieves workloads off of engineers. Vblock transforms IT into being more efficient in regards to business needs instead of resources dedicated to hardware maintenance. Removing this overhead of resources allows IT staff to focus on new ideas and moving the business forward.
Think of a Vblock procurement the same as buying a chassis full of blades. Instead of buying 10 or 14 different 2U servers, the blade offers the same performance in a smaller footprint and less network maintenance. A Vblock takes that another step further and completely aggregates IaaS in a box.
Where is this going to be in 5 years? More players and new alliances will make it's way into the market. VCE was formed by the best names in the industry, VMware, Cisco, and EMC. Each one of these companies are best of breed technologies in their related field and continue to innovate. As we all know, first to market is key piece to survival. In 5 years, as integrated stack purchases become an standard solution, we will see the hypervisor war start to die off. Why would new companies align themselves with a solution that they perceive as below par? In addition, once a company is tied with an integrated stack, it will develop new technologies that will benefit from having a complete stack. We will also see new 3rd party products come to market such as Nimsoft for Vblock by taking advantage of APIs to create more monitoring and management solutions for stacks. The API concept will increase throughout the whole stack. Almost all Vblock components have an open API concept that allows more management functionality. Backup vendors can take a complete backup of a Vblock (VMs, Blade Configs, Storage Zoning and all) and to be able reload that backup on a new Vblock. 1 set of backups for your whole infrastructure, pretty crazy, but it's only a matter of time. As bandwidth becomes cheaper and cheaper, multi-site VM hosting using Vblocks and VPLEX will create a more highly available cloud.
In the upcoming years, procurement of new equipment will be in the form of integrated stacks. It's a very logical way for IT to easily scale. I saw a great tweet yesterday from @clou
dbuzz "Nothing against ITIL guys, but one sent me a "cloud" doc that includes a 2-4 hour process window for a new VM request.. #notcloud". Being able to rapidly deploy and provision VMs for app scaling is a real thing, and being able to quickly scale out new hardware without the need to do everything in the first paragraph makes that a reality. In addition, more companies will be moving to a hybrid cloud or even public cloud approach. To meet the demand of all these new customers and resources, Service Providers will need to be able to quickly move on acquiring more infrastructure. Vblock meets those demands. There is no other product on the market like a Vblock. Remember, Vblock is a product that arrives completely assembled and ready to go. Reference architectures are merely directions on a piece of paper to mix together an infrastructure, which is no different than everything in the first paragraph.
I've heard this before... "If my manager buys a Vblock, do I have to worry about my job?". Well, who is going to manage the Vblock? There is always going to be a need for people in IT infrastructure that understand how all pieces operate. Virtualization changed job roles way before Vblock made its way to market. The server, storage, and network people all started learning how the game was changing and became VMware Admins. As the industry moves closer to unified fabric, a dedicated engineer to one specific asset is going to become rare. Vblock simplifies this even more with Ionix UIM. UIM orcheastrates a Vblock or multiple Vblocks to divide, provision, mask, zone, boot, manage, and monitor. Encompassing many roles into a single window. Diverse roles will still be needed in the infrastructure. Having a single "Cloud Architect" doesn't mean they understand what's going on behind the storage scenes and doesn't mean they can also administer the entire vSphere Cloud Director environment. Job roles change and being in IT, it's all about adapting to those changes.