Issue link: https://cp.revolio.com/i/324708
© 2014 Raritan Inc. These are just a few of Raritan's Innovations. Go to www.raritan.com/InnovativeSensors to get our new eBook 5 Reasons Environmental Sensors are used in all Modern Data Centers High Outlet Density Secure Locking Outlets Clear Printed Labels Innovators share two traits: uncommon vision and an unwillingness to compromise when a more effective solution is possible. It's those qualities that lead the world's most innovative companies to Raritan for their Intelligent Rack PDU solutions. It's Why the World's Most Innovative Companies, Innovate with Raritan. Meet the World's Most Powerful USB Port • Wi-Fi networking • Quick Setup • Cascade Multiple PDUs • USB Webcam Support • iPad Integration Make Sense of It All with Always-on Sensors • Outlet Level Metering • Billing Grade kWH • Full Environmental Monitoring • See Trends & Alerts • Improve PUE Color Coordinate for Greater Control • Multiple Colored PDU Options • Easily Identify Power Feeds • Secure Lock Cables • Faster Troubleshooting • Reduced Downtime RAR13034_PPD_Innovator_DCJournal_05192014.indd 1 5/20/2014 11:02:48 AM 12 | THE DATA CENTER JOURNAL www.datacenterjournal.com have constructed a basic carbon-nanotube computer. Speculation ensued that it could be a prototype for computers that are faster and less power hungry than their silicon- based brethren. Carbon nanotubes face a several hurdles in terms of their use in a processor. Semiconductor manufacturing processes are extremely delicate owing to the tiny dimensions of the components. Fabricat- ing nanotube-based processors has been thwarted in large part because of imperfec- tions in alignment of these carbon strands. Another problem is that some nanotubes, rather than behaving like semiconduc- tors (which can be "switched" on and off ), behave like wires—that is, conductors. ese sorts of defects, even if they occur relatively rarely, can make the product nonfunctioning. e Stanford design man- aged to avoid these problems, but it only involved about 200 transistors. Today's chips, however, oen contain billions of transistors, representing a 10-million-fold difference in scale. At the larger scale, ran- dom defects could make carbon nanotubes unworkable. Like its flat counterpart, graphene, carbon nanotubes offer interesting proper- ties, but they also suffer some of the same problems. At a nontechnical level, they are a solution to a problem that has not been clearly formulated. e traditional semiconductor industry has so long domi- nated computing that as the end of Moore's Law draws nigh for silicon, researchers oen appear to be grasping at any nearby straw to find the next wonder material for computing. is approach isn't neces- sarily wrongheaded, but a breakthrough may need to await more context—both technical and relative to what business and consumers need. what if siliCon is it? e acceleration of technological development, particularly in computing, has given many people the impression that progress in this area is the norm. But numerous factors—not the least of which being economic—all but guarantee that the pace of change will slow. In particle physics, for instance, experiments require multi-billion-dollar equipment, which only governments can afford (particularly given the returns, which largely comprise only physicists getting their jollies, and perhaps a few obscure publications along with a popular headline or two). Naturally, a similar stall in technology development in other fields—such as computing—could likewise follow. Other alternatives to traditional silicon-based computing have been pro- posed, ranging from integrated photon- ics (which doesn't necessarily involve a wholesale replacement of silicon, however) to biomolecular approaches. But what if silicon is really the end of the road for the wild ride dictated by Moore's Law? First, innovation will not stop completely. Even if Moore's Law fails, the computer industry could still see impres- sive innovation—it will just be tamer and harder won. Instead of expecting process- technology innovation to address all prob- lems, designers will need to focus more on architecture development and better soware efficiency. Second, businesses, individuals and governments might get a chance to tackle the social implications of the technology, such as the potential abuse of big data and widespread surveillance. Contrary to the optimism of some, tech- nology doesn't necessarily yield a net ben- efit. It is always a mixed bag, but the more informed and thoughtful the population, the greater the chance that technology's use will be more good than bad. Perhaps the necessary condition to finding a silicon replacement is waiting for silicon's progress to run its course. As long as Moore's Law continues with traditional semiconductors, the economic incen- tive for finding a replacement are mini- mal—particularly if that replacement isn't immediately far better in some respect. For instance, a business probably wouldn't replace a perfectly serviceable data center with years of life le just for the satisfac- tion of employing a new IT architecture— again, unless that architecture delivers benefits that justify the cost. ConClusions e waning of Moore's Law has natu- rally led to a search for silicon's successor. Quantum computing, although perhaps the most hyped possibility, is also the least likely near-term technology. Develop- ment of a scalable system—to say nothing about programming it—is years off at best. Whether the strange phenomena of the quantum world can be made to serve macroscopic purposes in an economi- cal manner is debatable, D-Wave's claims notwithstanding. Graphene and its close cousin, carbon nanotubes, exhibit characteristics that may lend these materials to practical computing applications. ey have already been built into prototype devices (transis- tors or simple processors), but again the question of scalability arises: can they serve in the same large-scale manner as silicon, which enables billions of transistors on a chip? Here, too, even if the material offers tremendous physical benefits, its use could easily be derailed by economics if, for instance, the cost of manufacturing or the cost of products is too large to justify the performance advantage. If silicon turns out to be irreplace- able, all is not lost. Innovation will continue, but it will eventually fall short of Moore's Law, making performance gains rarer than consumers and businesses have become accustomed to. e focus may simply switch from process technology as the chief source of advancement to archi- tecture (such as heterogeneous comput- ing) and soware (more-efficient use of hardware). Society may also get a chance to take stock of all the concerns that arise from decades of frantic change, which has led to big data, ubiquitous surveillance and the Internet of ings, for instance. But the decline of Moore's Law may end up being anything but dramatic: as the PC market has demonstrated, eventually consum- ers get their fill of computing power. e average home user would be unlikely to feel a pinch were advancement to stop completely. Only time will tell whether the successor to silicon is a material cur- rently under research, a material yet to be discovered, some new computing model altogether or simply nothing at all. Moore's Law may or may not stand whatever the outcome, but innovation will continue regardless. n