Heard of the Software-Defined Data Center? Here’s the Software-Defined Blog Post
June 17th, 2013 By: Industry Perspectives
Brian Reagan is vice president product strategy and business development, Actifio. He was previously CTO of the global Business Continuity and Resiliency Services division at IBM Corporation, responsible for the technology strategy, R&D, solution engineering, and application development for all global offerings including cloud services.BRIAN REAGAN
Each calendar year can be easily associated with a “tech meme.” 2011’s Cloud gave way to 2012’s Big Data. 2013 is nearly halfway over and it’s clear that this year’s meme is “Software-Defined”—specifically in my line of work, the “Software-Defined” Data Center.
I’m not suggesting that these secular trends aren’t / weren’t valid. Nor am I saying that these are not transformational forces that will radically alter the way we conceive, design, build, and run IT for the next several decades. They’ve already started to have a significant impact in companies large and small.
My beef is that as each of these tech waves emerged, every tech company under the sun – and even some non-tech companies (try googling “Refrigerator Cloud” for some terrifying stories of home-tech gone bad)– felt compelled to hook their message wagons to the trend, valid or not. Sorry, I’m not buying it when a 30+year software concern – born in the era of mid-range computers, COBOL, and loads and loads of tape – declares their solutions as “cloud-ready.”
Hype Helps Marketers
For the industry gorillas, it’s an easy play as they create these message waves. He who creates, must own. And so it goes. For the startups and emerging players, it can be perceived as a survival skill – adapt your product positioning to the latest industry message trend, or become irrelevant.
Software-Defined was picking up steam on its own, but truly broke through when the Software-defined networking company Nicira was gobbled up by VMware. I’m waiting for the first Software-Defined Software concern to enter the market.
So, now it’s the Software-Defined Data Center. All caps are required when we’re talking about Important Technology Ideas. Messaging and hoopla aside, this is the nirvana that we’ve been talking about for the entire 23+ years I’ve been in this industry. We’ve invented service management frameworks to automate, we’ve developed orchestration stacks, we’ve eliminated friction in the provisioning of resources, we’ve created open standards for system communications. Cloud promised a lot of what we’re seeing in SDDC messaging, yet we’re still not there.
Looking Beyond the Labels
This is certainly a sophisticated topic that has far-reaching architectural implications. But, it does boil down to three simple ideas.
1. Application Centricity
Business Applications are the center of the IT universe. The development and operation of these applications is the CIOs primary concern. Boards of Directors don’t concern themselves with LPARs or LUNs. They care about EPS and, by extension, the applications that drive revenue, manufacturing, customer satisfaction, etc. Perhaps a more appropriate term, then, would be “Application-Defined Data Center”?
2. SLA-Based Automation
The IT resources required to support applications – compute, network, storage, management – must be aligned to the nature and criticality of the specific business application. SLAs create the common language between IT and the business / application-owner. Of course, SLAs are not paper documents filed away for “checkbox” compliance. Automation of these SLAs is critical, in order to orchestrate the necessary resources to deliver on the agreed-upon terms of performance, availability, protection, resilience, retention, etc.
3. Heterogeneous Support
There are many products in the market today who’ve built an entire value proposition on the premise that a data center runs 100 percent virtualized (typically on VMware). Sure, there are some born-in-the-cloud businesses who have the luxury of a homogeneous infrastructure, but I’ve not met a single customer in the past several years who doesn’t have at least 25 percent of their applications running on physical systems. And, these are not corner-case, bought-in-the-90s-and-can’t-end-of-life-yet applications. These are bread-and-butter, mission-critical systems running state of the art versions of operating systems and databases. So, any Software-Defined (or Application-Defined) Data Center must be able to comprehend the entire application portfolio, not pick-and-choose those that map to a limited interoperability matrix.
I’ve left off some old chestnuts like “scalability” and “manageability” as I believe we’ve reached a tipping point in IT – in the face of the data volumes common in even the smallest business, these are table stakes for ANY IT-oriented product.
I’m hopeful that we can move quickly past the Software-Defined hype, and to the more important discussions regarding the architectural inhibitors to progress – particularly those brittle, legacy approaches to core IT processes. At the end of the day, shouldn’t it be the user that defines the data center and not the software? The end user should define how and when it wants its data. Maybe in 2014 we will see a transition to discussions around the “Application-Defined Data Center.” Until then, I’ll see you at Cloud-Based Coffee Machine. Maybe we can talk some Big Data while we’re there.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.
You are spot on with the application-centric view and the need for SLAs. The interesting thing that I think is starting to get traction now is that these SLAs need to be expressed in application terms, not networking terms, for them to be useful. Asking an application guy to understand networking on networking’s terms is not really reasonable. They need to express application requirements (or intent) based on their world.
This abstracted view of SLAs and policy will be necessary on top of the good work happening with SDDC and SDN already.
Mike Bushong (@mbushong)