Even as technologists, we rarely think twice about the behind-the-scenes technology that enables some of our favorite consumer cloud technologies – Gmail, Facebook, Twitter, iTunes, etc. While we all know that there is world-class scaling and PhD created algorithms, our focus is shifted to the value they provides in our day-to-day lives; how they solve problems in how we work, live, play and interact. And while there are discussions and arguments over aspects of the underlying technologies (HTML5 vs. Flash, POP3 vs. SMTP vs. IMAP, iPad clients vs. Blackberry clients, etc.), the infrastructure is rarely ever discussed. It just works. It just scales. Problem solving first, technology second.
As we began delivering VBlock to our customers and partners, several questions came up in almost every technical discussion.
- · “Why can’t I pick and choose what goes into my VBlock?”
- · “Couldn’t I just build this myself, without the VCE guidance?”
- · “Doesn’t this lock me in to a vendor-specific solution?”
These are all fair questions. As technologist and engineers, we’re trained to look for every way possible to improve efficiency, to improve performance, to solve the problem at hand. But the mindset of “seeking perfection” creates some challenges in today’s Data Centers, especially if one of the goals is a more flexible environment that delivers Private Cloud functionality.
Before I elaborate on that thought, let me back up and explain some of the thinking behind why we offer VBlock and how it answers some of the questions above. At the foundation of VBlock is a set of world-class technologies spanning Management, Virtualization, Computing, Security, Network and Storage. Technology that not only is leading in market-share, but also leading in performance and innovation. VBlock takes these technologies and delivers them as a unified solution, balanced for performance across multiple use-cases and workloads. With the combination of VMware, Cisco and EMC (including RSA) technologies, plus the open ecosystem of application technology partners (Microsoft, SAP, Oracle, RedHat, etc.), it’s rare that we ever hear customers or partners question the technical foundation. In many cases, it was the foundation of portions of their existing networks.
The second foundation of VBlock is our belief that technology alone will not completely help customers on their journey to the Private Cloud. There is a people aspect, a process aspect, and an organizational aspect. These elements, which drive the 70-85% of operational expenses (OPEX) of the Data Center, will need to adapt and change to meet the pace of the underlying technology. The technology, driven by virtualization, is already blurring the lines between how we manage every aspect of the Data Center. Where does the network stop or start; who should manage storage for virtual machines (VMs); are our security policies still valid when VMs are mobile? Technology challenges indeed, but also organization challenges and barriers that will ultimately decide if your Data Centers will ever operate as an efficient Private Cloud.
Once you’ve decided that a Private Cloud is one of your Data Center goals, VBlock is the first step in aligning the technology and organization strategy. By presenting solutions that enable holistic thinking, at the solution-level, VBlock helps customers and partners focus on solving business problems first. VBlock helps align the technology groups, allowing them to address how they will deal with the technology shifts that were occurring before VBlock came into the market.
Let’s get back to the mindset of “seeking perfection” and infinite tweaks. The perfect infrastructure to deliver services for a breadth of business requirements. It’s an admirable goal. But business trends and cycles are difficult to predict. It’s compounded by data growth that can only be comprehended if it’s expressed in analogies with multiple commas (billions of Tweets per second?). So I’d argue that the perfect design is nearly unobtainable. How can I say that when there are so many outstanding architects and engineers building Data Centers? Because if such perfection was possible, there would never have been such an enormous market for the first wave of VMware server virtualization and consolidation of boxes running at 5-10% utilization. This isn’t a knock on architects and engineers; they were using the tools at their disposal in the physical world. They designed for expected performance, for software bugs, for operational errors and other things that occasionally caused that 5-10% CPU utilization to jump to 99% and frustrate every end-user. They didn’t have VMware HA or DRS. They didn’t have VMotion or SVMotion. They didn’t have UCS Service Profiles, UCS Extended Memory, Fully Automated Storage Tiering (FAST), Fast Cache or Enterprise Flash Drives (EFDs). Technologies that automate around failures, capacity-planning challenges, and hardware/software abstraction. We now live in virtualized world.
So now we have better tools, greater automation, and the ability to design “pools of pools” that deliver dynamic capacity that allows architects to address typical business needs as well as unexpected spikes or variations. These tools are embedded in Vblocks, and are delivered in a way that allows those pools of resources to be managed as a unified solution. The alignment of technology and operations to deliver Private Cloud services.
It’s a different way of thinking about delivering services to the business. It requires a slightly different way of looking at problem solving, but we believe that it offers a superior model for both technology and operational efficiency.