Industry Perspective is a regular Data Center Journal Q&A series that presents expert views on market trends, technologies and other issues relevant to data centers and IT.
This week, Industry Perspective asks Brian Reagan about the software-defined data center and its relevance to today's facilities. Brian is vice president of Product Strategy and Business Development for Actifio and has over 20 years of experience bringing solutions to market. He most recently served as CTO for IBM’s Business Continuity and Resiliency Services division.
Previously, Brian was CMO for Xiotech, being responsible for reshaping the company’s brand and product family. Before Xiotech, he was CMO for Arsenal Digital Solutions, an early leader in the cloud-based data-protection space that was acquired by IBM. Brian spent eight years in senior-level global marketing and strategy roles at EMC, which he joined after leading an online services development team at MCI Telecommunications. He started his career in IT consulting, serving clients primarily in the financial services, pharmaceutical and government sectors.
Brian holds a Bachelor of Arts from Bennington College in Biological Sciences and a Master's in Business Administration from George Mason University.
Industry Perspective: What is a “software-defined data center” (SDDC)?
Brian Reagan: SDDC is a more efficient vision for managing IT, versus today’s labor/tool-intensive model of standing up infrastructure silos by application. In an SDDC, the infrastructure resources of a data center are decoupled from the applications that run the business. Automation of the service levels required by each application—performance, availability, resilience, retention and so on—is done by uniquely combining and orchestrating the underlying resources. It goes without saying that standardization and simplification are core attributes of an SDDC.
SDDC is an important stepping-stone towards a true application-defined data center (ADDC), but it is still infrastructure-focused. ADDC is application-down, versus infrastructure-up, in which virtualization is also applied to the operational processes that drive the application-defined SLAs.
IP: Is “software-defined” in this case related to virtualization?
BR: VMware certainly is advancing the idea of software-defined as a logical extension of its virtualization offerings—beginning with compute and network, and extending to other layers of the IT stack including data and storage. There’s no doubt that virtualization is the cornerstone of an SDDC, but there is more to the story. Automation or orchestration is critical, as are security and governance. Management, particularly in the definition and application of SLAs, will also play an important role in making SDDC, and ultimately ADDC, real.
IP: How does a SDDC differ from a more traditional data center?
BR: If you looked at the physical plant, you’d be hard-pressed to tell the difference. You’ll still see racks filled with switches, servers, storage, green lights blinking, fans blowing and so on. The outcomes driven by an SDDC, however, are radically more efficient and effective. For example, an SDDC would require far fewer people to operate, even as volumes grow.
IP: What are the benefits of an SDDC approach?
BR: The executive summary is “do more with less.” SDDC meets hyper-scale requirements without commensurate infrastructure and operating costs. Consolidation of infrastructure drives space and energy savings. Standardization and simplification result in streamlined processes and faster time to productivity. SDDC also promises higher levels of resilience, as resources are applied or redirected to respond to interruptions. Time to market for new applications is also reduced, as resources can be made available when development and QA need them.
IP: Many new technologies and data center strategies become overhyped and then fail to meet expectations. Where do SDDCs currently fall in this regard?
BR: It’s still very early in the SDDC lifecycle, so the hype far outpaces adoption. That said, SDDC is evolutionary, particularly from the last wave of hype around cloud. So, many of the core building blocks are maturing, particularly in the areas of network and compute. Still, these technologies are still very much focused on infrastructure. They’ve not yet made the leap to comprehend and respond to application requirements. This will be a growing area of focus and investment over the next decade.
IP: As has happened with the cloud, are some so-called software-defined solutions not really software defined?
BR: With any hype cycle, there are companies and technologies that continually rebadge themselves. The telecom/SP industry is a great example. In the late 1990s, hosting companies renamed themselves application solution providers (ASPs). Then, they delivered SaaS in the 2000s, until they rebadged again under the cloud moniker. I’m sure their marketing departments are working overtime on how to incorporate SDDC into their company profile.
There’s no doubt that we’ll see the same thing play out in software and infrastructure. Credibility will be challenged when a company with a primary product that was built 20 years ago using a core functionality that hasn’t changed drastically since then decides to declare it is “software-defined.”
That’s why it’s so important to agree on a definition of what software-defined really means. Then, businesses can evaluate vendors on a more objective basis.
IP: Is an SDDC an all-or-nothing approach, or can data center operators implement portions of their facilities using software-defined methods?
BR: Today, there are varying levels of maturity in the tools and technologies required to move to SDDC, so it’s by default an incremental over-time model. IT leaders will undoubtedly focus first on areas of great pain and/or inefficiency, then apply additional elements as business need requires.
IP: Is SDDC a strategy for all data centers, or is it best for certain applications?
BR: At a strategy level, SDDC is universal. The outcomes driven by an application-centric approach are transformational. Reality, however, dictates that certain applications and/or platforms will be easier to incorporate into a SDDC strategy than others. For example, the applications being developed today, which are born in a virtual machine using state-of-the-art frameworks, will be first movers. On the other hand, I don’t envision VMware or any other SDDC vendor taking great pains to focus on applications running on IBM iSeries.
IP: How do SDDCs relate to big data?
BR: SDDC and big data are certainly competing for attention in the hype wars. In all seriousness, SDDC should be a critical enabler of big data. There are many challenges enterprises face with big data including integrating diverse systems, moving to real-time insights and identifying the “source of truth” for select data. Most of these challenges are directly related to the inefficiency of infrastructure silos. In a world where applications call the shots, the ability to gather, analyze, synthesize, and report across the enterprise will be accelerated.