Bandwidth and Network Speeds Exploding with Hyperscale Deployments
Data center infrastructure is rapidly transforming to a more cloud-friendly footing, and this will no doubt put an end to many longstanding design and configuration principles.
One of the more widely suspected targets is the rack server, which, although useful in hyperscale settings, is still seen by many as out of sync with the modular, scalable direction that the data center needs to adopt in order to remain relevant in the emerging dynamic data environment.
But recent research is putting that notion to the test. According to MarketsandMarkets, the data center rack server market is expected to nearly double over the next five years from today’s $22 billion to more than $40 billion. As with a number of other technology sectors, North America will remain the largest market, although the highest growth is due to come from the Asia-Pacific region. To be sure, the report predicts some dramatic changes in the capabilities of rack servers, particularly when it comes to power efficiency and high-performance capabilities, but in the end the segment is expected to not only survive the cloud transition, but thrive.
Of course, this is by no means an indication of smooth sailing for today’s commercial server vendors, even when it comes to the lucrative hyperscale market. Companies like Facebook are more than willing to deploy rack configurations in their modular data centers, but these are usually stripped-down commodity machines sourced directly from original device manufacturers (ODMs) in the Pacific Rim. The company is attempting to standardize its approach through the Open Compute Project, which offers hope that smaller firms will select rack devices through traditional distribution channels, but these will still be simple devices without all the bells and whistles that usually enhance hardware profit margins.
It’s hard to imagine that the top server vendors are not aware of this trend, however. As EnterpriseTech’s Timothy Prickett Morgan notes, manufacturers like Dell are eager to tap into the growing hyperscale market with solutions that stress density and scalability but with enough cachet to rise above a simple commodity solution. The new PowerEdge FX line, for example, is designed to meet the form factor and efficiency requirements of scale-out architectures while avoiding the networking and management hassles that accompany blade or microserver configurations. As well, it incorporates 1.8-inch SSDs and a DAS storage module linked by an I/O Aggregator that does away with multiple network interfaces, mezzanine cards and switches.
Ultimately, the decision to deploy racks, blades or some form of hybrid will fall to workflow requirements and business needs. As systems distributor Tuangru notes, racks require a lot of cables, but these are easy to manage within most standard racks. As well, racks are noted for a high level of failure containment due to their isolation and ease of replacement. On the downside, they tend to require a lot of cooling, which drives up operating costs, and their larger form factors tend to eat up valuable data center real estate.
So it seems clear that rack servers still have a healthy shelf life in the data center, but like all hardware they will need to evolve into sleeker, more efficient versions in order to provide key support for the modern data user.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata, Carpathia and NetMagic.
