Demystifying SDDC

MUMBAI, INDIA: Raghu Raghuram, EVP & General Manager, Software Defined Data Center, VMware, speaks about the developments in SDDC and how the technology is transforming businesses, in a free-wheeling interview with CIOL.

What’s the update on SDDC?
The software-defined data center (SDDC) is part of a larger architectural shift that is happening in the industry and that is being increasingly referred to as the mobile cloud architecture. The billion endpoint devices are fundamentally bringing changes back in every part of the IT chain from the device if you work backwards all the way into the application. That is the macro story that is happening in the industry.

So this is causing a change in applications because the delivery model has changed. It is causing a fundamental change in end-user experience. It is causing a fundamental change in network architecture. It is causing a fundamental rethink of security and it is causing a fundamental change in the back-end data center as well. In fact, going one step further, it is causing a fundamental change in how businesses think of their business models.

You have heard the well known cliché by now that software is eating the hardware world. Most businesses are becoming software businesses. It does not matter whether you happen to be in logistics or banking. So that is the sort of the Uber picture that is happening. SDDC has proven to be the linchpin of how the back-end data infrastructure has to evolve. Whether you are operating a cloud, whether you are modernizing your in-house data center or whether you are doing something in between.

That SDDC is eating up the hardware business is given. How are enterprises adapting to this change?
In the case of SDDC, software is eating hardware. The fundamental realization that Google and Facebook and others came to several years ago is the only way they could deal with the scale that they were being forced to deal with is turn infrastructure into a software model and deliver it on top of industry standard hardware.

That was the only way they could get to the scale at a reasonable economics. And that was the only way they could automate the rate of change that they wanted to introduce. For them, the most critical thing was how fast you are introducing new capabilities; as you introduce more changes, customers expect more innovation.

Who is adopting SDDC?
If you think about server virtualization as the most well developed of these software-based data center technologies, we have over 5,00,000 customers and the average customer is of the order of about 70 percent virtualized. In places like North America, Western Europe and Australia, it is probably about 80 percent or so. In certain other parts of the world it is lesser. In India, it is 24 percent. So server virtualization is most deployed of the software-based technologies and that is a market of around 5,00,000 customers. All other technologies are in the low single digit percentage adoption compared to that. Software-based storage of these 5,00,000 customers and this obviously starts from global one to mom and pop stores in all sorts of geographies. Not all of them are candidates for all the software-based technologies.

So the modernization is happening today in I would say in three class of customers.

One, are the truly large multi data center customers that are finding that their operational Capex costs are pretty dramatic. The second is companies that are technology enabled-and banking, logistics, software development, hi-tech, oil and gas, you could name a whole plethora of these where fundamentally software is a key differentiator. Software is how the products get delivered and so you need a software-defined data center. They have started adopting that. The third of the smaller companies there are I would say born-digital companies.

Those are the markets in which SDDC is getting deployed but it is early days. I would say overall worldwide we are still in the low thousands of customers. High hundreds, low thousands of customers.

Would one of the reasons for low adoption also be the ROI factor?
It is remarkably easier actually. Yes, because at the end of the day you still have to run a data center. So there is a tangible comparison if it is running as a purely hardware-based approach and when you start a software-based approach. And in fact this was the original comparison that propelled server virtualization. So we were able to go to a customer and say instead of running ten servers; run ten virtual machines on one server.

I can do the same thing for storage now. The average customer deploying software-based storage saves about 60 percent on Capex. A customer deploying network virtualization saves a similar number on network firewalls and switches. So the dramatic deflation of data center costs can be very tangibly proved with SDDC architecture.

What is the value proposition of SDDC that can be offered to customers?
When we walk into customers, we offer five value propositions.
One is anywhere from 40-60 percent savings in Capex that in turn translates into fewer administrators and doing things faster. The third is being able to enforce security at a per application level which no other technology can do, even if you have money you will not be able to do that. The fourth is reduction in the cost to make your data center 99.99 percent available for all applications. Today it is done only for the critical databases. And the fifth is being able to deploy applications continuously from–collapsing the time it takes from a developer writing code to the time it goes into production.

Each value translates into the economic benefit. Customers used to adopt one or two to begin with, and they tailor the economic case that way. So that is really what we see in the marketplace.

Why is adoption then a challenge?
Because this is a pretty dramatic shift in the people and process. So the traditional data center organization is you have a server, network and storage teams and they all have their favorite brand of hardware vendor just like you and I have our favorite brand of consumer technologies and they go to irrational lengths to protect that.

It is not irrational. Their technical competence is very closely associated with whether I have CCIE or whether I have EMC certification or whether I have. And the IT organization is built that way. The IT leadership is built from these sorts of models. With SDDC, you have a software infrastructure team and that is principally the team that is an intermix of competencies between networking, storage and it is all software. And companies have to make that shift and that is a tougher shift than just technology.

The second is the way you serve the internal customer changes. If you are an internal customer and I am the data center guy–previously, in a hardware-defined data center model, you would write up a help-desk ticket or something. It will make its way and then six months later things will get satisfied because the server guy will go and deploy the server, then the ticket will be handed over to the storage guy and he will deploy some storage, etcetera.

Whereas SDDC is truly like a cloud model, companies that adopt SDDC have a front-end and a back-end team. The front-end team is almost like a set of product managers that works with the business units to assess the demand for certain types of products. Then they work with the back-end team to create the infrastructure for those products. And then they put it in a service catalog and you can go and order it off just like an app store.

So that is a very different model of how you serve your end-users and it is a people’s change, it is a process change and that does not happen readily. And in fact it only happens when a forward thinking CIO says, look you guys shape up or I am going to the external cloud altogether.

That is really the glaring factor more than technology or technology readiness.

Tomorrow: Beyond SDDC