Data CenterData Center Architecture

Collision on Big Data: Dealing with the New Dynamics

We’re witnessing the start of explosive growth in Big Data, along with the headlong pursuit of the Big Data opportunity by technology vendors from four camps. This trend of tech titans colliding on Big Data—highlighted in my last post—is deeply changing the dynamics of the data center market.

Defined broadly, Big Data and its infrastructure requirements takes in everything from consumer trends like video streaming over mobile broadband, to enterprise-class analysis of data gleaned from smart equipment, sales trends, or consumer digitization behavior. Big Data is enabled by technologies including virtualization and blade servers, and is driving the need for larger data centers and Cloud-based services. As noted in the last post, Web giants are building massive data centers, or what I call “motherships” for Big Data as part of the trend.

I think almost everyone recognizes a major shift is taking place with Big Data, so the question really becomes, what do we do about it? What do we as providers and users of data center solutions need to focus on to gain the most benefit from the changing dynamics?

Well, there are multiple factors to focus on, including:

  • The move from capital expenditure (CapEx) to OpEx.
  • Big Data equals high density and has to be globally available.
  • Traditional production and disaster recovery data centers plus many more now in an enterprise.
  • The titans seeking dominance need to be global, efficient, and consistent.
  • Quality of service will now be network driven.(bandwidth)
  • Energy management and DCIM will help set the dominant players apart.

Let’s examine some of these factors in turn.

First, these early days for Big Data have come at a time when most companies are under tight capital budgets due to lingering effects of the global financial crisis. So there is the need for massive capacity, but few companies can lay out the CapEx to build their own data centers. Fortunately, there are major technology players such as colocation data centers providers, or Colos, as well as IT giants looking to transform around Cloud-based solutions so that your more average enterprise can meet their needs on an OpEx basis.

For data center solution providers, this move from CapEx to OpEx is a big change. Rather than the old days of selling equipment to companies building their own data centers, much of the growth is coming from helping the technology titans who are building the motherships, supporting regional hubs, and need reliable network infrastructure. So data center solution providers need to focus on things like helping Colos expand globally, helping Telco companies protect and modernize their mobile broadband hubs, or helping Colos or Web giants build highly efficient, modular data centers that simplify the integration of IT systems, data center physical infrastructure (DCPI) and other systems like security or building management.

At the same time, data center providers still need to help more traditional end users transition to this new era, such as by helping them consolidate to fewer, more efficient data centers, while helping them protect their connections to the Cloud. So in this move from CapEx to OpEx, there is both the need to smooth the transition for more traditional end-user companies, while helping the Big Data titans strengthen their global reach, infrastructure, and data center efficiencies.

Second, it’s important to remember that Big Data means high-density data centers. These can be challenging to run, not only because they have denser, hotter, more compute-intensive loads, but because of virtualization, which means rapidly shifting loads. This in turns calls for updated DCPI solutions, such as modular data center pods for rapid expansion, containment cooling, and data center infrastructure management (DCIM) software that easily links with virtualization engines to coordinate DCPI with the dynamic IT loads.

One example of how Schneider Electric has adapted these needs is to create standard integration between our DCIM tools and virtualization platforms from vendors such as VMware and Microsoft. Our solution for Colos, for instance, benefits from this type of integration, as well as DCIM visibility into factors such as cage and rack space to efficiently meet multi-tenant requirements. In effect, data center providers are going to have to be more like partners to the titans by helping them with these sorts of killer apps.

The third dynamic to consider is the global reach needed to support Big Data. For instance, we’re seeing big Colos expand into new countries by buying up data centers or building new ones. Telcos that used to be national in scope also are expanding internationally. Web giants—in addition to building a handful of “motherships” for Big Data—are building smaller regional hubs to support their Big Data goals.

As these titans expand globally, they need a data center partner who itself has global reach and can help them roll out operations on a consistent basis. This need might play out, for example, by finding a standard way to monitor power usage effectiveness between data centers, or by simplifying the integration between DCIM and IT tools.

The fourth dynamic to consider is energy management and operational efficiency. For the companies vying for Big Data dominance—the Web giants, the Telcos, the Colos, and IT companies—their data center facilities and network infrastructure need to be highly available, efficient, and secure. These goals tend to elevate the importance of energy management and DCIM for these companies, because they have to set themselves apart to their end users on factors like uptime, cost, quality of service and security.

To support Big Data delivery, it’s not just conventional data centers that need to be efficient and highly available, so does the related network infrastructure. We are seeing tech and Web giants building out edge points of presence (PoPs) to support Big Data and Cloud computing, while Telcos need to modernize their hubs to support mobile broadband. In this digitized, Big Data era, these network assets have become more like data centers that require best-in-class DCPI components to ensure reliability.

So where does all this leave data center solution providers? Well, we’re becoming Big Data infrastructure partners and advisors, helping titans build out and reliably operate their infrastructure, while also helping more traditional enterprises transition into the new era.  Now more than ever, we’re not selling boxes. We’re about partnering with both the titans and the more traditional end user companies to ensure they achieve success with their Big Data business models and goals.


3 Responses
  1. Mostafa Bandar

    Very interesting blog and well presented. Thanks Dave.
    It would be interesting at some point to also have a view of Service offer & deliver expectations from such customers and market drivers.
    Regards
    Mostafa

    Reply
  2. Tim Walters

    Great article, David. Big Data is definitely an area that is growing exponentially. It will be interesting to see how some of these large data centers can manage their infrastructure and expand as necessary.

    – Tim

    Reply

Leave a Reply

  • (will not be published)