Edge Computing Base
First, the basic concept of edge computing
In the era of the Internet of all things, all things Internet of Things includes not only the environment, including lower interconnection between things, also includes an Internet context-aware, more computing power and perception of persons and things. All interconnected to the physical network based on integration of intelligent network, and the synergy between all the visual function.
Sensors, smart phones, wearable devices and intelligent home appliances and other equipment will be part of all interconnected, and generates huge amounts of data, and now the network bandwidth and computing resources cloud computing model can not process the data efficiently.
Cloud computing cloud using a large number of resources to process data, but in all network environments, a conventional cloud computing model does not effectively meet the needs of all Internet applications, mainly due to the following:
(1) directly to the end edge device transmits data to the cloud mass, resulting in network bandwidth and computing resource waste load.
Privacy Protection (2) the traditional computing model of the cloud computing model will be an important challenge facing the interconnection of all things cloud architecture.
(3) Energy most things interconnect fabric edge device node is limited, and the GSM , the WIFI large wireless energy transmission module.
For more than a few problems, the use of edge device already has computing power, application service program to migrate all or part of cloud computing tasks from the center to the edge of the device-side implementation, which will help reduce energy consumption.
An edge calculating means calculating a new model calculation is performed at the network edge, the calculation of the specific data comprises: the cloud service line and upstream of all connected services. Calculation edge "edge" refers to any computing between a data source to the cloud from the center of the path, storage and network resources. Transmission request between the terminal apparatus and the center of the cloud are bidirectional. Only a request from a network edge device and the center of the cloud content services, but may also perform part of computing tasks, including data storage, processing, cache, device management, privacy protection.
However, in terms of a certain angle, the edge calculation model is a distributed computing system, and has a flexible management synergistic implementation, and real-time processing environment of heterogeneous characteristics. Edge computing model also includes the following key elements:
(1) an application / service functions can be divided
(2) Data may be distributed
(3) the distribution of resources
Second, the edges of the key technologies computing
(1) calculate migration
Cloud computing model, computing migration strategy is to migrate computationally intensive tasks to sufficient resources in the implementation of cloud computing centers. But massive edge equipment and vast amounts of data transmission is an important factor affecting their computing power, so edge computing model to calculate migration strategy is to reduce the amount of network traffic data for the purpose of migration policy, rather than compute-intensive tasks migrate to the edge device to perform .
Calculating an edge calculation migration policy at the network edge, the edge devices collecting vast amounts of data generated or partially or preprocessing operations all calculations, filtering unwanted data to reduce transmission bandwidth.
(2) 5G ICT
5G data communications technology is the development of next-generation mobile communications core technology of a new era. 5G high bandwidth, low latency, low power consumption and other advantages not only used for communication between people, but also for communication between objects, things and objects. At present, China for the 5G offers three technology scenarios: embB (enhanced mobile broadband), MMTC (massive machine type communication), uRLLC (ultra-low latency reliable communication). Which eMBB scene mainly for VR / AR grade high-bandwidth demands of business; MMTC scene mainly for the wisdom of the city, intelligent transportation needs of high-density connector business; uRLLC main scenarios for unmanned drones and other delay-sensitive business.
After the edge device capable of handling some or all of computing tasks and data filtering garbage sensitive data, or the need to upload the data to the intermediate or final cloud data center. 5G technical solutions will be necessary to reduce the mobile terminal edge data transmission delay.
(3) new storage system
Calculating an edge with a strong demand for real-time processing and data storage, the storage system more computing edge latency characteristics, large capacity, high reliability. Data calculated feature edge has a higher timeliness, and relevance diversity, the need to ensure a continuous edge data storage and preprocessing, and therefore how to efficiently store and access uninterrupted real time data , the storage system design is the need to focus on the edge calculation Concerns.
High-density, non-volatile storage medium low power, low latency and high read and write speeds will be deployed at the edge of large-scale equipment, non-volatile storage medium facing challenges in the limbic system.
(4) lightweight kernel and libraries
Network edge by different vendors exist in the design and production of massive edge devices that have a strong thing heterogeneity and performance parameters vary greatly, so the application is deployed on the edge of the device very difficult. Virtualization technology as the preferred solution to solve this problem.
More resource-constrained edge devices need to support libraries and kernel lightweight, consume fewer resources and time to achieve better performance.
(5) Calculation edge programming model
Third, the edge of computing and cloud computing
Edge computing and cloud computing compared to calculate the edge not intended to replace the cloud, but complement and extend cloud computing, for mobile computing, networking and other computing platforms provide better. Edge computing model requires powerful computing cloud computing center of mass storage capacity and support for cloud computing also need an edge device edge calculation processing of massive amounts of data and privacy of data to meet the real-time, privacy protection and reduce energy consumption and other needs . Edge computing architecture: end device ---- edge ----- cloud three-tier model.
Fourth, the edges of the large data calculated
V. edge computing summary