From edge to 5G, Inspur’s battle for the beachhead

The word "edge" suddenly became extremely popular in the past 2018.

From an industry perspective, Microsoft, SAP, and Amazon have successively announced their plans to deploy edge computing to achieve connectivity from the cloud to the edge, making the market competition for edge computing very crowded.

The rise of edge computing is closely related to the deepening of cloud computing. When cloud computing first emerged, all terminal data was transmitted to the cloud for processing and calculation, and then transmitted back to the user's terminal, thus forming the Internet's business model. However, with the advent of the Internet of Things era, many specific application scenarios have very strict requirements on latency. For example, if driverless cars rely entirely on the cloud, the efficiency will inevitably not meet the requirements.

Therefore, the concept of edge computing was formed, and many new market players were born. Inspur is the most typical one among them. It is their beachhead landing that laid the foundation for the smooth implementation of edge computing in the 5G era.

Edge computing will drive three types of players onto the stage

In fact, the concept of "edge computing" itself is not a new term. As early as 2003, CDN service provider AKAMAI cooperated with IBM to launch the earliest "edge computing". Objectively speaking, if Amazon's launch of AWS in 2006 is regarded as the origin of cloud computing, it is earlier than the time when cloud computing was proposed.

In my opinion, the IoT world composed of cloud computing and edge computing is particularly like an "octopus". Cloud computing forms the brain of the octopus, and edge computing is the tentacles of the octopus. Obviously, you must have a brain before you can command the tentacles. So it is not an exaggeration to say that the vigorous development of cloud computing has promoted the development of edge computing. The combination of "edge computing + cloud platform" is also recognized as the future.

Another driving force is 5G. In the current network architecture, due to the high-position deployment of the core network, the transmission delay is relatively large and cannot meet the ultra-low latency business requirements. This determines that the endpoints of 5G services cannot all be on the cloud platform at the back end of the core network, and mobile edge computing just meets this demand. On the one hand, edge services run on terminal devices, with faster feedback, solving the latency problem; on the other hand, edge computing sinks content and computing power, provides intelligent traffic scheduling, and can achieve business localization.

This means that technologies such as autonomous driving, VR, AR and IoT have sufficient capabilities for implementation, and new experiences in scenarios such as high-definition video, telemedicine, remote work, and distance education will benefit first.

It is foreseeable that the development of edge computing will inevitably lead to a series of business opportunities. Providers of edge computing infrastructure, represented by Inspur, mainly provide edge servers and edge gateways. For example, at the just-concluded World Mobile Communications Conference MWC 2019, Inspur released the first edge computing server NE5260M5 based on the Otii standard. This product is designed for 5G and can undertake 5G application scenarios such as the Internet of Things, MEC and NFV, and is suitable for the physical environment of the edge computer room.

Second, because edge computing and CDN have many technical similarities, CDN service providers will have new development opportunities; third, as cloud computing sinks to the edge, the importance of data security and network security will be self-evident, and security vendors will have new growth opportunities. In the 5G era, we also expect these vendors to become the main players on the edge computing stage by landing on the beach.

The challenge of edge computing comes from the “edge” itself

Gartner stated in a report that centralized traditional data centers will no longer exist in the future, and IT infrastructure will be distributed and ubiquitous.

There are several key prediction figures in this report, such as: edge computing infrastructure will grow the fastest. By 2020, 50% of large enterprises will include edge computing in their plans, and more than 50% of industrial IoT analysis will be performed at the edge; second, by 2022, 50% of enterprise production/processing data will not be in the data center or cloud; third, by 2025, enterprises will close 80% of traditional data centers, etc.

What these numbers mean is that edge will be the fastest growing part of infrastructure in the future.

But, what is the current situation?

Previously, Internet giants have exposed their own data centers, all with standardized cabinets, powerful cooling systems, and standardized network cabling, which is a visual enjoyment. However, edge computer rooms are often a different story. As far as I know, the conditions and environment of most operators' edge computer rooms can be described as "bad". Not only are the spaces small, but the heat dissipation and ventilation conditions vary greatly, and the operation and maintenance are extremely difficult.

To use a vivid metaphor, if the cloud data center is like a hotel, then the edge computer room is like a roadside guesthouse. In summary, the servers in the edge computer room actually face these challenges.

For example, the harsh deployment environment: many edge computing environments are very scattered and the environmental conditions are very complex. The traditional server design framework is difficult to meet the diverse requirements of the space. For example, the AI ​​server supports the wall-mounted design. The 220V electricity used in residential buildings can only withstand a certain amount of power. This poses a great challenge to the power consumption of the server. In order to ensure business stability, the difficulty that the edge server must overcome is the environment.

The second is the structural design of the server: the power supply and fans of standard servers are placed at the back, but the space in the edge computer room is often relatively small, so some customers require a front window design suitable for operation and maintenance, with the power supply and fans placed at the front. This places high demands on the customized services and delivery capabilities of edge servers.

The third is daily operation and maintenance: Traditional data centers have dedicated maintenance personnel to perform maintenance, but edge computer rooms require targeted design strategies for equipment maintenance due to their dispersed geographical locations.

The fourth is security: We know that data centers are closed operations with clear security guarantees, but edge computer rooms are scattered across the country and may even be prone to theft or vandalism. How to take data security issues into consideration during product design is a new topic for edge device providers.

The rush for edge computing is also a rush for the 5G era

It is not difficult to analyze what an edge server should look like from these challenges.

First, in order to meet the space utilization and deployment challenges, the size of the edge server should be as small as possible; second, the physical environment of edge deployment is very poor, so the edge server must meet the reliability design requirements such as heat resistance, high temperature resistance, humidity resistance, and dust resistance. The customization requirements of edge servers are extremely high, and flexible and efficient on-demand customized production and delivery conditions are also necessary.

Judging from these conditions, the NE5260M5 launched by Inspur at MWC is definitely an edge server with extremely strong applicability.

[[278075]]

First, in terms of product volume design, the chassis size of this product adopts the telecommunications equipment standard rather than the server standard. It is 2U in height, 19 inches in width, and 430mm in depth, which is only slightly more than 1/2 of the depth of a traditional standard server. It can be directly mixed with telecommunications equipment and deployed on the communication center rack.

Secondly, in terms of environmental adaptability and reliability of the product, NE5260M5 has a lot of adaptability designs for edge computer rooms, such as wall-mounted design, which allows NE5260M5 to be directly hung on the wall without a rack. At the same time, it is also suitable for edge data centers with simple environments, such as high temperature resistance, dust resistance, corrosion resistance, electromagnetic compatibility, earthquake resistance and other harsh conditions.

Third, in terms of application scenario scalability, NE5260M5 supports 2 next-generation Intel scalable processors, 16 DIMMs, 6 PCI-E slots, 6 2.5-inch hard drives, and 2 M.2 SSDs. Under limited space constraints, it still has very ideal scalability and can meet the needs of various edge computing applications such as video surveillance, smart manufacturing, and the Internet of Things.

In terms of agile innovation, Inspur's JDM model can ensure that Inspur can quickly deliver targeted designed products for different operators and different edge deployment environments.

As the industry's first OTII server, Inspur's beachhead landing in the 5G era actually has many practical significance.

We know that OTII is a server open project jointly initiated by China Mobile, China Telecom, China Unicom, Intel, Inspur and other companies. The project was launched in November 2017. The goal is to form a deeply customized, open standard, and unified server technology solution and prototype product for the operator industry for telecommunications applications. The fundamental problem it wants to solve is the challenges of standardization, environmental adaptability, and maintainability faced by traditional servers in the core business application scenarios of communications. As an important member of OTII, the launch of the first OTII server by Inspur will also win a first-mover position for Inspur in the edge computing track in the 5G era.