With the growing popularity of Hyper-Converged Architectures from companies such as Nutanix, SimpliVity, Scale Computing, and now VMware’s VSAN, we continue to see a growing trend and a transformation of the Data Center. Software which provides Policy driven Management, Automation, and Efficiency today are what virtualization brought to the Client/Server era years ago. It is and will be the driver of the Data Center today and years to come.
Now that the primary focus is moving away from complex Hardware Infrastructures with many management points, the focus has shifted to truly converged solutions with intelligent software providing single point management and services at the compute, storage, and network layers, and all within lower cost commodity hardware. The true value is in the software and the intelligence of that software. Software that can be layered on many types of hardware, more-so commodity to keep costs in check, and provide Policy based MGMT at the Virtual Machine (App) Level and also has the intelligence to monitor, and make changes as needed to keep up with performance demands, (Tiering), provide operational analytics and provide a robust local and remote protection scheme will be the victor in the Data Center Infrastructure War.
This leads me to the title of this post, Web Scale. We have been hearing this from the Hyper-Converged market for a bit now, and it’s becoming more commonplace. Let me first point out that not too long ago, I didn’t think this was the right term to use as Marketing for the Hyper-Converged players out there because I was looking at the “Scale” in Web scale to mean massive “Google-like” scale out nodes of commodity servers and I came to the realization that “Scale” in Web Scale means much more than that. When taken out of context as i did, my first thought was why an organization that wants to focus on mid-market and even have products that are promoted for ROBO environments, use the term Web Scale? Let’s face it, most customer outside of Upper Enterprise (Private Cloud), and Service providers do not need Scale in a massive scale out sense. They need lower cost entry to market, grow as they go type architecture and in this sense, Hyper-Converged architectures still deliver and then some.
While it’s true Organizations like Google, Amazon, Facebook, etc needed massive scale out while keeping the performance gains:
It’s also true the the time to implement and stand up these nodes is very rapid. This not only due to the converged nature of the infrastructure, but also the intelligent software that allows the seamless addition or deletion of nodes and automatic balancing to bring optimal performance. In this sense, Scale is defined as a quick deployment model.
The icing on the cake with these architectures or the “Mojo” I like to call it is the Software. It’s the Policy based management and automation that deliver additional operational savings.
I deploy VMware products on Semi-converged architectures today, but the problem and what I still hear from the customer is “I love the solution you and your organization have implemented Mike, but now I have all these interfaces I have to go to to manage the environment” which can be daunting to some, and a pain in the ass to others.
This is where these new truly converged Products can help. They can deliver a simple management model while still providing reliability, performance and services that a semi-converged platform can offer, yet easier, and with a somewhat “set it and forget it” approach. Yes this will be an operational shift, yes it may be difficult to deliver that message, but IMHO, I feel it’s a necessary message that our customers need to hear and will only benefit from and at the end of the day, that’s what it’s about.