Micro-server market: The kingdoms of the unified or the Jiangshan

If the 2013 micro-server market is a pre-war exercise, then 2014, with the release of 64-bit ARM servers, the battlefield of the micro-server is officially unveiled.　　Micro-server is the server subclass of Data Center Application subdivision, which is suitable for the rapid development of the current lightweight, parallel load, compared with the standard server has lower energy consumption and cost, in recent years, has gradually been the focus of data center market. Previously, the chip dealer did not give the market enough attention to the reason, the market was not ready. A few years ago, the Internet was not as developed as it is today, and cloud computing and big data are just rhetoric, and even the data centers have low energy needs that are not as strong as they are today.　　Now that the market is ready, the micro-server is naturally pushed to the cusp. x86: Need enough attention if x86 a few years ago enough attention to low power server market, now I am afraid also will not appear x86, ARM and MIPS dispute situation. Previously, whether Intel or AMD, for the Low-power server market strategy, is roughly the introduction of the server processor Low-power version, or simply the desktop processor, and there is no clear independent of the micro-server product line.　　Although Intel's first generation Atom was later used on the HP Moon Server, Atom was originally designed to be desktop rather than server. Finally, before the 64-bit ARM release, Intel's second generation Atom Chip "Avoton" came out, this time, Intel early action, is the real. Without arm, Intel may also follow its own steps to a step-by-step low power market layout because Intel has no rivals in the data center--not to overlook Amd,amd's Hao Dragon X, which is a fairly competitive product, almost entirely beyond the first generation of atom,　　But it's not enough to do it in a low-power server market. Relative to arm and Mips,x86 has more advanced technology and manufacturing capacity, has a better compatibility and versatility, but also has unmatched complete ecological chain and market base, in the data center, x86 has accounted for more than 90% of the share. In other words, if x86 has more appropriate products, why should users risk compatibility and platform migration to choose another architecture? Therefore, x86 is not lack of technology and market, but for the micro-server enough attention.　　Even with a challenger like arm, x86 has enough strength to become the dominant force in the micro-server market. ARM: The market still needs to accumulate three years ago, just when x86, almost unified data center, ARM came, at that time, Facebook intends to use the ARM server in the data center. This message sparked two diametrically opposed views, one for arm's nose: arm to do a good job in the old line of the mobile end, for the data center, ARM is still inexperienced, especially with a lack of 64-bit computingand software support for the mishap.　　Another part of the applauded: arm than x86 cheaper, more power-saving, what reason not to choose it? At the time, though with low power and low cost advantages, but compared with the mature x86,arm really lacks in the enterprise market experience, after all, the mobile market and the data center completely different, need to learn and accumulate; In the process, arm still lags behind x86 many, the production capacity also completely cannot catch up with intel such chip big crocodile　　On extensibility, standardized x86 chips are easier to expand and have greater scalability; more crucially, in systems and software, the lack of sophisticated operating system support, the poor application software, and the lack of programming tools support the X86,arm. So, although arm will aggressively enter the data center of the slogan has been shouted for three years, but the real use of ARM server is still very few manufacturers, but Facebook, Baidu and several powerful independent research and development capabilities of large internet companies.　　It can be said that 32-bit ARM server is just a test of water products, arm need to accumulate experience in the enterprise-level market, but also through the move to attract partners eyes, the real kill recruit, is about to come out in 2014, the 64-bit arm. Perhaps because of the success of the 32 arm test data center, perhaps arm's low-power, low-cost features really on the chip manufacturers and server OEM taste, ARM has now won the support of many partners, 64 arm is ready. In the recently held "arm is in" technical conference, including AMD, Calxeda, Applied Micro, Marvell and other manufacturers have announced the progress of the 64-bit ARM chip.　　Applied Micro, for example, showed the 64-bit ARM test sample, and Mitac also demonstrated a prototype of a 64-bit ARM architecture Micro-server that runs standard Linux systems and KVM virtual machines on this prototype. AMD and HP will be arm into the data center of the important partner, ARM lacks the enterprise-level market experience and market base, and AMD has a wealth of server chip design experience, Hewlett-Packard has a mature enterprise-class market and strong appeal, with these two support,　　ARM's march into the data center is much faster. Although many manufacturers are actively preparing for war, but many manufacturers are still watching.　　A well-known domestic it manufacturers, the general manager of the server department has admitted that arm chip products and applications are not yet mature, the current release of a lot of products, but the market is very small, and the manufacturer is not intended to be the arm of micro-server products, but to wait until the micro-server market "clearer" time. In other words, despite the 64-bit arm is bullish, got a lot of chip manufacturers and server OEM support, but arm has not as x86 as all server manufacturers of "necessities", ARM micro-server market from matureThere is still a distance. MIPS: A short-term fear of nothing as in early September, the UK's largest CPU and semiconductor production company Imagination Technologies announced that the MIPS technology will be extended to servers, mobile devices, smart wearable equipment and other fields.　　Imagination bought MIPS at a cost of 100 million dollars last year, and the goal was ambitious to cover all the CPU-related areas. Imagination's business model is like ARM, open MIPS licensing to partners, and MIPS features similar to arm, using RISC architecture, in power consumption has excellent performance. But because of the similarity, MIPS also faces the same problems as arm: 64-bit computing, compatibility, operating system and software support, enterprise-class product and market experience, ecosystem building ... Wait a minute.　　More unfavorable is, and in the mobile market a single big arm compared with MIPS now almost disappeared, but also in the application of MIPS few areas, how to give partners confidence, regain support is a difficult problem.　　At least for now, MIPS are not bullish, and imagination has frankly said that MIPS wants to make sound in a competitive market, at least for a few years.　　The kingdoms of the unified or the Jiangshan? Considering that MIPS will take some time to enter the micro-server market, in fact, the micro-server market players only x86 and arm, to say "The Kingdoms" is too early. and arm relative to mature x86, also is a new person, also not be the same heavyweight player.　　Although ARM has a number of chip manufacturers and server manufacturers support, but just like x86 into the mobile market is blocked, arm want to enter the enterprise market is not easy, or lack of enterprise market experience and perfect ecosystem, but also the accumulation and maturity of the market. However, the market demand is diverse, in the enterprise market Advantage obvious x86 want to kill arm in the cradle, it is not an easy thing, after all, arm has attractive low power consumption and low-cost features, this is the current spread of the data center is a great temptation.　　There is demand, natural market, in fact, arm into the data center market, in the final analysis or market demand. Perhaps, in the 2014 we will not see the Battle of the micro-server market, but undoubtedly the market will be more and more prosperous and exciting.