详细内容

Intel has led the trend toward shorter interconnection of optical connections

时间:2017-07-13 15:32:03   来源:   作者:   点击:

  What should the future of device interconnect look like? Is it a myriad of cables that are intricately connected, poorly laid out and very expensive?
Now, Internet and mobile devices and consumer electronics products, the common problem facing is how to do a smaller, thinner, lighter, more grab an eye, such as thinner laptop computers, smart phones, TV, etc. This pursuit of thin and thin involves device interconnection.
  Status quo of equipment interconnection
  Now there are so many different connectors that cause a lot of problems. Many connectors not only take up space, but many connectors themselves are large enough to increase the volume of the corresponding device. For example, a laptop has a video connection hole, an audio connection hole, an Internet connection hole, and various slots. May really need a laptop computer connection holes are one or two, but after all, these connectors are installed, so it occupies a very large space, also use to bring trouble to the user.
  In many cases, we use too many cables and too many types of cables. Open a closet in your home and you'll find that there are all kinds of cable wires in it, and every function is different. When the user needs to find a cable, it may take a long time, so the user experience is very poor.
  On a technical level, our ability to achieve connectivity on cables is very limited. Copper technology is getting closer and closer to its limits, and its costs are very high, and the cost of using these cables is very high.
  For the manufacturers who want to make ultra-light TVS hanging on the wall, they can take up a lot of space if they have slots and sockets. Consumers often have to turn the TV down, from the wall and think about what which hole should be plugged into the cable, and each equipment of cable and different, such as multimedia equipment, computer equipment, communications equipment cable is different, different devices have different needs. It's getting harder and harder to put all the joint holes and slots on one machine, because there's not enough space. This raises a very serious question: how to design devices in the future, and how will the connectivity of future devices be realized?
  Short connections are the trend
  Before we talk about connectivity, let's talk about wireless. In addition to wireless transmission, the future of very important technology development is optical fiber transmission. Progress in this area has been slow, but it has not stopped. In the long run, optical fiber transmission is our first choice. High bandwidth, long distance and large capacity transport advantages determine that optical path transmission is a long-term solution.
  Many problems in the past have been solved wirelessly. But in the future, Intel China research institute and other Intel research department, technology department, the important technical direction is light path optical fiber transmission technology, it is very important transmission technology in the future.
  Now that the application of copper technology is approaching the limit, you may ask: what is this limit? This limit doesn't have to be determined on a certain number, for example, we can increase the amount of transmission by 1 meter per time. Now that the USB3 is no longer a 5m-long connection line, it has been shortened to 2~3 meters, and the connection line shortening is a trend.
  At the same time, the cost of the copper interconnection approach has come close to the limit, and we hope that the optical connection will be able to connect the devices in a very low cost way. In the future, we can find a very cheap way to connect light. In telecommunications, optical connectivity is sure to be the preferred way to connect.
  Optical interconnection is the future direction
  Optical connections are often used for long-distance interconnection, such as 50 km or more. Because it is sometimes required to make optical fiber in deep underground, the requirements for the test environment are very high. So now, the cost of optical connectivity is still very high, and you might have to spend $100,000 to do that. Only when the price is in a reasonable range can consumers really accept fiber transmission.
  Now, only when we find a very cheap way of optical interconnects, to ensure that consumers can accept, so that consumers from handheld devices to a variety of higher end equipment can adopt the way of the Internet. So we're going to measure the cost, not just a couple of dollars per port, but not a couple of hundred dollars, so you can get end-to-end connectivity.
  Two things could happen in the future: one is that users of different industries use optical interconnection to realize interconnection between devices in their own devices; Second, there are multiple connectors on a device, such as a connector with a monitor on the monitor, and a versatile I/O connector. If we move in the second direction, there will be a completely unnecessary market segmentation. Because if each market, each device has its own unique connector, it will not only increase the cost but also be very cumbersome to use.
  Intel has realized that optical connectivity is the future, and we have seen a general convergence of optical connectivity. The effect of scale can be achieved through optical connections, and it is very convenient for consumers to use. So what are the key technical requirements to implement?
  First, there should be cheap, fast light connections; In addition, we need to reduce the number of connectors, so we need to figure out which links can support different kinds of traffic. In this way, a single connection can support a wide variety of traffic. We have to solve a problem: we don't need a a connector, we need is a very good way of exchange, exchange between multiple devices can realize direct traffic, this also can further reduce the cost. Ideally, there would be a connector family, such as a server family, a connector family of handheld devices, and a very limited number of connectors, and an open one. Equipment may not be the same today. In the future, no matter what device is invented, these connections can be supported. This ensures that various devices are well connected. This is the essence of the LightPeak project.
  We are working with the industry to study this technology, and it is a single, compatible optical architecture, which is our goal. From a practical standpoint, LightPeak provides users with such a system, all have the same capacity, jack if your cable can plug into a socket, can realize the function you want to and transmission capacity.
The architecture that LightPeak can achieve is first with a merged I/O transport layer, which is not designed to replace an existing protocol, such as a way to change the way you transfer pixels and images. The people who built this architecture fully understand the way people do hard disk data transmission. We don't want to compete with existing industry technologies. We just want to create a common architecture, platform, and support for a variety of different protocols.
  The technical details are described below. Intel started with a basic program, and everyone says we're going to crawl. The plan is first and foremost by a router carry a variety of protocols, in the long run, our aim is to ensure the good fusion, so as to achieve the performance of high speed and low cost. In the past few years, Intel has been working on silicon optics, which can only be achieved with such technical capabilities.
  Intel's opportunities and challenges
  In February 2005, Intel released its first hybrid silicon laser, and has since released some other more critical hybrid silicon lasers. Silicon can bring heat, but not light, which creates problems. The hybrid silicon laser solution is a collaborative project between us and the university of California, Santa Barbara.
  In addition, we have achieved remarkable results in the field of radio and television detectors. A lot of the work we do now is to achieve the traditional capability of carrying capacity and technology. In the case of radio and television, the bandwidth we get is the highest in the world. We can say that our radio and television detectors are far more capable than other existing technologies.
  If you can achieve the above technologies and want to break into the market, you must consider two points: one is the low cost and very high speed (100Gbps) photoelectric connection mode; 2 it is cost-effective, ultra-high speed (bits per second) solution, in the center of the cloud computing, big data, the technology is very necessary, can make the data in a powerful system for rapid transmission.
  Intel is also facing some challenges, and these challenges are largely due to energy efficiency, and we are working on that. We know that energy efficiency is important for every system. Another challenge is encapsulation. The problem of encapsulation is how to encapsulate the fiber, there are a lot of solutions, but what is the best solution, a big part of it is energy efficiency.
Intel thinks that in the next few years, it will be possible to move light technology to the mass market, and the time is ripe. And it will change the way connectors work. A big part of it is technical challenges, you know how to use it, you have to do basic research, and you have to do engineering innovation. All solutions need to be standardized, and the industry needs to work hard to get our new vision implemented as soon as possible.
 

Copyright©2009-2010  AnhuiLongCableGroupCo.,Ltd.  All rights reserved ICP备案:皖ICP备08002904号
Contact phone number:0553-6867189/6618036/6618008   The true:0553-6865288   E-mail:ahlacable@163.com 
网站制作及维护:巢星网络   To address:Jiangjiawan bridge, new avenue, wuhu county, wuhu city, anhui province   Mail make up:238371

皖公网安备 34022502000116号