Issue link: https://cp.revolio.com/i/324708
8 | THE DATA CENTER JOURNAL www.datacenterjournal.com Moving lots of Fiber! Needs BIG DATA ©Copyright 2014. Free * $20 coffee card...cablesys.com/promo * With $500 order of fiber jumpers. You just spent hundreds of thousands of dollars building a BIG data center. Now comes the time to connect the servers and switches and you go to an online supermarket to buy fiber jumpers? NOT SMART! Cablesys does not sell supermarket grade jumpers. For decades, we have been manufacturing high performance, high speed, data center grade fiber jumpers. Our mission is to assist network integrators with implementing data center projects and we make sure they don't worry about a single connection. Thousands of fiber jumpers will light up without a hiccup. We can pre-label and pre-bundle to save you time and make your life easier. Best of all, we are 40% less than most big brands. Give us a try by calling us now at 800.555.7176, you will be surprised how easy and cost effective it is. We Can Help! mpr-0464 - DATA CENTER JOURNAL JUNE 2014.indd 1 5/14/2014 9:10:31 AM i t's happening all around us, driven by new innovations that are building upon the cloud foundation of the third platform. If the first wave cloud was focused on user focused and consumer applications, this next wave is being driven by machine-to-machine and industrial applications. is is where the "Internet of things" and "smart machines," among other developments, are quietly changing the way we control and interact with devices in fundamental ways. And this revolution isn't taking place in the data center—it's happening at the edge. In fact, with the tremendous growth in the cloud and at the edge, the traditional client-server data center model shrinking. why foCus on the eDge? Edge computing is all about pushing processing—including applications, data and services—away from the core of the data center to the very place where the work is being done at the extreme of the network. is could be a telematics system for autos, a machine in a manufacturing plant, or even a consumer appliance. Why push computing to the edge? First, because some applications require extremely low latency. Running these ap- plications in a cloud that may be located hundreds of miles away isn't practical. ey require real-time control or machine-to- machine interactions with extreme avail- ability. Bringing server resources closer to the devices doing the work—embedding them, in some cases—is the best way to achieve this. Second, moving to the edge is about reducing network traffic. As "smart" devices and machines proliferate, so does the data they require and produce. Placing dedicated computing resources locally can dramatically reduce network traffic. Of course, edge systems will ulti- mately need to be seamlessly integrated with the cloud and other services. is will require remote monitoring—especially for applications in those remote locations where there is no IT staff—and data gener- ated at the edge to be gathered, consolidat- ed and analyzed. So "performance critical" processing is occurring locally at the edge, with centralized cloud storage and pro- cessing, making data available to mobile devices for anytime, anywhere access to trend information. is edge infrastructure looks very different from the computing resources in the data center or the cloud. In the data center the typical server lifecycle is five years and in the cloud its two to three years at most with constant "care and feeding." But things are different at the edge. ere, server lifecycles may be seven to ten years. Edge servers must be easy to deploy, easy to use, and easy to manage, with remote monitoring. With no IT staff onsite, they must be extremely reliable. While three nines (99.9%) availability may be typical in the data center, edge infrastructures must deliver five-nines (99.999%) availability— no more than 5 minutes a year of un- planned downtime—to support business- critical applications. three eDge exaMples So what do these edge server applica- tions look like? You may be surprised. Walk into a popular national hamburger chain and you may see an unusual-looking so drink dispenser. e Coca-Cola Freestyle machine lets consum- ers select more than 100 different beverage choices—from sparkling water to caffeine- free so drinks in a wide range of flavors (including colas, of course)—all with the push of a button. ere's even an iPhone app to help consumers locate the nearest machine, wherever they go. It's currently the world's coolest beverage machine and each one relies on an edge server to make it run. is server must be optimized to a particular task very quickly. And it gath- ers a wealth of information on consumer preferences that it makes available to the "mother ship" via the cloud. Smart. Perhaps the greatest area of growth for edge servers is in manufacturing, where machines just keep getting smarter and smarter. Increasingly, manufacturing machines are being designed to communi- cate with each other, coordinating complex processes without human intervention. is machine-to-machine communication relies heavily on edge servers optimized for low-latency processing of very specific functions day-in and day-out without a break. ese servers, in turn, gather valuable information that can be used to further optimize production processes and proactively spot potential trouble areas to help anticipate and prevent costly downtime. Another fascinating example of edge server intelligence keeping things on track can be found in the latest generation of trains. To improve rail safety, the U.S. Congress passed in 2008 new regulations mandating the implementation of new positive train control (PTC) technology across the U.S. rail network by 2015. PTC technology is intended to monitor and control train movements to ensure proper train speed and separation in order to min- imize the risk of collision and to improve track worker safety. Key to PTC is the deployment of on-board equipment that continually calculates a train's speed to a target speed dictated by the train's location and track profile. Such a speed control unit relies on a database of track profiles linked to a real-time navigation system. And that requires a high-speed server optimized for high availability processing in a harsh en- vironment. With more than 20,000 trains operating on 65,000 miles of track across the U.S., that's a lot of data processing. aVailaBility is non- negotiaBle To meet the stringent availabil- ity requirements in these demanding deployments, engineers designing edge deployments must take advantage of the latest technologies to ensure minimal downtime for edge servers—including soware-defined availability for virtual servers. Indeed, IDC has pointed to "edge- optimized" virtual servers and storage as a major trend for businesses seeking optimal remote location deployment strategies for the extended enterprise. 1 is is just the tip of the edge iceberg. Even as the cloud continues its explosive growth, just as rapid is the expansion of intelligent processing at the edge in every dimension of our lives—even including the wearable devices people are increasingly using to improve their "measured lives." Wherever there's an opportunity to im- prove the efficiency of a function with real- time processing, while gathering valuable data to aggregate and analyze, you'll find an edge server humming away happily. It's just another facet of the "always on" world that's quietly changing everything. n about the author: Dave LeClair is the Senior Director of Strategy at Stratus Technologies 1. Villars, R., "Enhancing Business Value with an Edge-Optimized Virtual Server and Storage Deliv- ery Solution," IDC White Paper, June 2013.