Shulou(Shulou.com)06/02 Report--

For server manufacturers, whoever takes the lead in edge computing will lead the future.

The reason for this is that thanks to the rapid development of artificial intelligence, Internet of things, big data and other technologies, as well as the official commercial use of 5G, "things" began to become the center of information, generating more connections and intelligence based on "things". As a result, computing requirements extend from the data center to the edge. IDC predicts that at least 40 per cent of Internet of things-related data will be stored, calculated and processed at the very edge by 2019.

The edge computing market is like a blue ocean. Compared with the traditional data center market with mature products and stable market pattern, it has great potential for growth in the future. MarketWatch predicts that the global market for marginal computing will reach $21 billion by 2023, with a compound annual growth rate of 14.2 per cent. At the same time, edge computing is also in the stage of "barbaric growth", the application scenarios of many industries still need to be explored, product forms are varied, and application ecology needs to be established urgently.

There are very few edge computing servers in the current market. Because, different from the traditional server, it tests not only the scene exploration ability and product landing ability of server manufacturers, but also the construction ability of edge computing ecology. Lenovo, which has a deep foundation in server and computing, obviously wants to seize the opportunity in the marginal computing market. Therefore, Lenovo is the first to launch its first compact edge computing server ThinkSystem SE350 tailored for edge computing and AI in the market, and hopes to seize the opportunity in the edge computing market by paying attention to ecological construction.

ThinkSystem SE350: edge computing servers born for the edge will be a new species in the computing market. Because edge computing is really different from traditional data center server products in terms of product architecture, use environment, application ecology and so on. If you continue to use traditional data center server products in the edge computing environment, it will be difficult to meet the various needs of edge computing.

Zhou Tao, Marketing Director of Server products in Lenovo DCG China

At present, the main challenges of edge computing include the following four aspects:

First of all, there is a huge difference in the edge computing environment. Compared with the constant temperature and humidity environment in the data center, the edge computing may have high temperature, high humidity and high dust, and the current environment will also be different due to different environments, not to mention the differences in space, load bearing, wiring and so on. This differentiated or even worse environment has higher requirements for the design of edge computing server products.

Secondly, the ecology of edge computing applications is also a major challenge. Most applications are very sensitive to latency and cost, and delay is a key issue that must be considered. In addition, in addition to the need to understand the operation of edge applications and services in real time, it is also necessary to achieve cloud-edge collaboration and convergence of edge computing applications.

Third, compared with the data center environment, edge computing servers have no special care and access restrictions because of the differences in the environment. Therefore, the edge computing server not only needs to guard against potential network security problems, but also needs to consider how to prevent malicious destruction at the hardware level, and to ensure data consistency, high availability and no disclosure.

Finally, there is the operation and maintenance challenge of the edge computing server. Usually because of the deployment environment, the edge computing environment often lacks professional IT operation and maintenance personnel. How to achieve automatic and omni-directional unmanned operation and maintenance is also a big test of the edge computing server products.

"Edge computing server product architecture is really very different from traditional servers. It actually integrates all IT architectures, filters the data, then stores, calculates and processes it locally, and then collaborates and communicates with the cloud." Zhou Tao, director of server product marketing for Lenovo DCG in China, said, "the standard server has the advantages of large size, high power consumption, single network connection and high deployment environment, while traditional gateway devices also have problems such as insufficient local workload capacity, lack of redundancy and management capabilities, and high expansion costs."

Based on this, Lenovo launched the edge computing server ThinkSystem SE350 to fill the gap in the market. According to Wang Yunke, product manager of Lenovo DCG China Product Marketing Department, ThinkSystem SE350 is a 1U, half-width server with compact design, low power consumption, earthquake resistance, high temperature resistance and dustproof design, only 3.75kg, supports floor bracket, guideway or wall installation, supports multiple network access modes, and carries out redundant design on power supply and fan. And carried on the omni-directional authentication test to the environment, shock, vibration and so on.

Wang Yunke, Product Manager of Product Marketing Department of Lenovo DCG in China

"ThinkSystem SE350 has reserved 5G modules, and once 5G applications become popular, 5G access can be used." Wang Yunke revealed, "at the security level, ThinkSystem SE350 has strengthened the hardware and software. In addition, SE350 has excellent scalability, and GPU cards can be placed in expansion slots. Lenovo's automatic management XClarity can ensure that the remote operation and maintenance of SE350 is more intelligent."

As edge computing gathers manufacturers in many fields, and there is no clear standard definition of edge computing, the so-called "edge computing" in the eyes of a thousand people, it is not too much to describe the product form of edge computing in a variety of ways. In this regard, Zhou Tao believes that the edge computing server is not installed in the computer room, but in the edge space, which determines that the installation environment, conditions and shape of edge computing are completely different from those before. This will lead to the existence of different forms of edge computing servers in the next few years. " The appearance and volume of the edge computing server need to adapt to the application scenario. In the future, with the introduction of standards, legislation and industry standards for edge applications, edge computing servers will also embark on the road of standards. For example, now operators have begun to do OTII standards to regulate the size of some edge devices. "

Edge computing landing: ecology is the key edge computing landing in addition to products, ecology is also very important. Especially considering that the industry attributes and scene attributes of edge computing are relatively strong, different industries and different scenarios have different requirements for edge computing products and applications, and there are also high requirements for the construction of edge computing ecology. In the view of Yang Jie, director of server at Lenovo Research Institute, ThinkSystem SE350 is mainly aimed at application scenarios with heavy edges. "heavy edges are the most demanding scenarios, and the load to be processed in this scenario is almost equal to that in the data center, and often requires strong computing and storage capacity, including data preprocessing and analysis on the edge side."

Edge-focused applications are emerging in various industries, such as intelligent factories, intelligent transportation, autopilot, live events, and so on. "with the commercial use of 5G, there will be more and more heavy-edge application scenarios in the future." However, the landing of heavy edge applications still has a lot of challenges, and it is often necessary to consider and explore how to develop edge applications.

Yang Jie, director of server of Lenovo Research Institute

Yang Jie believes that there are three challenges in the implementation of current heavy-edge applications:

First of all, the hardware is not hard enough, the complexity and inferiority of the edge environment put forward higher requirements for the hardware, if the hardware is not good enough, it can not bear the requirements of edge computing.

Secondly, the software is not soft enough, and the trend of focusing on edge applications is that more and more applications are running on edge devices. How to make these applications run well and ensure the isolation and security between applications is a great challenge. Yang Jie revealed: "the isolation of virtualization technology is very good, but it is too heavy for the edge and takes too long; while the existing container technology can enable applications to start quickly, but without systematic isolation, there are potential dangers." Therefore, Lenovo believes that edge computing should provide the ability to launch applications in seconds, similar to that of containers and the isolation ability of virtualization, so that edge applications can be well supported.

The third is that the software and hardware are not close enough, and the current difficulty of edge computing application lies in the clear separation of software and hardware, and it is difficult for software and hardware to be optimized together. However, many edge computing applications, such as AR and VR, often need the integration of software and hardware to achieve good results.

Yang Jie revealed that in addition to investing in ThinkSystem SE350, Lenovo has also invested a lot of manpower in the corresponding technology research and development of the platform. " First of all, the hardware acceleration level, SE350 supports different types of accelerators, including WAN acceleration, GPU cards, etc., through the integration of hardware and accelerators to achieve edge hardware optimization. On this basis, Lenovo optimizes for different types of scenarios. For example, live streaming of events is an ultra-high-definition video application scenario that requires high storage and computing, while industrial manufacturing scenarios require a large number of connections and implementation responses, not only for edge devices with strong computing power. it also needs to be optimized at the operating system and network communication protocol layer. "

Zhou Tao said that edge computing products do not have very standard applications like other products, and the whole ecosystem needs to be built to allow product development at this stage. Lenovo has carried out comprehensive consideration and planning in edge computing from the very beginning. in addition to hardware architecture and products, colleagues at Lenovo Research Institute will focus on upper-layer applications to provide more perfect solutions in the future. Lenovo will also join more partners to provide a more perfect and reliable edge computing ecology.

Lenovo's Edge Computing Bureau in fact, edge computing servers belong to products with strong industry scene attributes. To some extent, it tests the manufacturer's ability to design products according to the industry scene, as well as the strong manufacturing capacity; in addition, the demand of edge computing for the application of the scene, it also drives that edge computing products must have a complete ecology in order to maintain long-term competitiveness. Zhou Tao said bluntly that once the edge computing server meets a suitable application scenario, the overall number of applications will be very large.

Lenovo's first Edge Computing Server: ThinkSystem SE350

"ThinkSystem SE350 is Lenovo's first edge computing server. Lenovo has a detailed and long-term product plan for edge computing server products. We will launch other models in the future." Zhou Tao revealed.

Obviously, Lenovo has an important investment in edge computing servers. In fact, the marginal computing market is also very important for Lenovo's future server business development. Under the circumstances that the data center market pattern is increasingly stable, edge computing, the largest incremental market, will be the hope for Lenovo server business to break through the market pattern. More crucially, Lenovo's edge computing server products ushered in a good opportunity for climate, geography and harmony.

First of all, the characteristics of 5G high speed, ubiquitous network, low power consumption and low delay determine that its commercial application will greatly promote the application of edge scenarios, which will greatly promote the demand and development of edge computing. Edge computing server is about to usher in its spring. Secondly, the Chinese market is currently in the leading position of 5G commercial, and many industries have begun to explore the scenario of 5G application. Lenovo's edge computing server will have a huge opportunity to display its talents in the Chinese market. In addition, Lenovo's edge computing server is not driven by a single business unit, but a collection of products created by multiple business units such as Lenovo DCG and Research Institute.

With the launch of ThinkSystem SE350, Lenovo has sounded the horn for the commanding heights of the marginal computing market. It is believed that with the gradual enrichment of Lenovo edge computing server products, Lenovo adhering to the product + ecological strategy will show its skills in the era of the Internet of everything.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.