Original Paper: Mao Y, You C, Zhang
J, et al A Survey on Mobile Edge Computing:. The Communication Perspective [J] IEEE Communications Surveys & Tutorials, 2017, 19 (4):.. 2322-2358 undertake Part blog limited capacity general level, meaning some sense out of the wrong place please correct me.
MEC listed first and cloud computing the difference table, the table from the paper.
Note: The so-called resource-constrained hardware performance is not so good
| null | GUY | Mobile cloud (MCC) |
|---|---|---|
| Server Hardware | Resource-constrained small data center | Large data centers (each performance server) |
| Server Location | Proximity wireless gateways, routers, and base station serving LTE | Services in the area of expertise, there is a special large room |
| Deployment mode | By the telecom operators, MEC providers, enterprises, and housing their own user-intensive deployment, simply planning and configuration | Deployed by a number of large Internet companies, such as Ali, Google. Require professional planning and configuration |
| Distance to the user | Small (a few hundred meters to several kilometers) | Large (probably across the country) |
| Frequency information return | Few, reduce network congestion | Often, it can lead to network congestion |
| Management System | Hierarchical control (centralized or distributed) | Control Center |
| Delay | 10 milliseconds or less | More than 100 milliseconds |
| application | Delay-sensitive (low latency) and compute-intensive applications (AR, autopilot, online interactive games) | It is not sensitive to delay and compute-intensive applications (social networking platform, online education, online business) |
Low latency
compared to the MCC, MEC has low latency, mobile energy-saving equipment, support context-aware computing, increase the advantages of privacy and security. And transmission distance, computing power, data corresponding to the rate, low latency characteristics for mobile services in three aspects: the transmission distance, calculating and communication delays.
-
A few dozen meters from the information transmission in a cellular network or dense small end (D2D) case, under normal circumstances will not exceed one kilometer. In contrast in the cloud transmission distance from an end user or from a core network node to the data center is generally from 10,000 meters to across continents.
-
Due to large-scale cloud computing capacity needed for complex operations, the performance of the device than the edge of the device is much higher. However, the edge device and compared to many users need to share the cloud server, calculating the time delay is reduced gap between the MEC and the MCC. And modern base stations have the ability to perform highly complex computing tasks. Call of Duty minimum requirements for example 13 compared to the configuration, a computing power of the base station 100 to 10,000 times the minimum configuration.
-
MCC requires information transmitted through a radio access network, a backhaul network, the Internet, in which the link control, routing, network management will result in additional delay. Because the MEC communications at the network edge, it will not be above the influence.
Studies have shown, MCC delay in 30-100ms range. Such a delay is unable to meet many applications (eg real-time online gaming, VR Sports, unmanned, these requirements within 1ms) claims. In contrast, using a simple protocol and MEC end of the transmission distance, such that the delay has the potential to achieve MEC haptic biological level, 5G and delay-sensitive applications.
Energy saving
due to the IOT device is limited energy storage device, but also want to do some complex and requires collaboration tasks such as monitoring, health monitoring. For these things device, the key things is how to design a battery energy deal with these large-scale repair and replacement of equipment of Things issues. MEC technical support uninstall computing tasks, which will extend the battery life of things devices. In particular, intensive complex computing tasks can be unloaded from the apparatus to the edge device things in order to reduce the loss of raw material can be networked device.
Context-aware
Another MCC MEC and the different characteristics that, MEC MEC edge devices and servers closer to the end-user, real-time tracking of user real-time information, for example, behavior, location and the environment. It is possible to provide context-aware services for users. For example, in the AR program the museum tour guide, it is possible to predict the user's interests through the user's location, at the same time push relevant content to the user. In CTrack system, you can monitor traffic, navigation, planning and personalized journey planning journey.
Improve security and privacy
in mobile cloud computing (MCC) system, the cloud computing platform is the remote center of big data platforms, such as Ali cloud, cloud Tencent, Baidu cloud. Since these cloud computing platform focused a great deal of user information, it is easy to become targets of criminals. And because the user data owners and managers are different groups, so easily lead to leakage and loss of data. However, the use of edge server closer to the user provides a promising approach to circumvent these problems. Due to the nature MEC server, distributed, small-scale, less centralized. Making it more difficult to become a target. When an application needs to exchange private information in the client and server, MEC is beneficial. For example, when companies deploy MEC, you can avoid the confidential data and information uploaded to the remote cloud, business managers no additional units will be able to manage the authorization, access control and different types of service requests.
