Bandwidth, Timeliness, Privacy: The thrust behind the edge tide
IDC predicts that as many as 50 billion smart devices will be connected to the Internet by the year 2020, and over 40% of the data in the future will need to be analyzed, processed and stored at the edge. These smart devices include smart phones, personal wearable devices, automobiles, MRI, smart street lights, steam generators, aircraft engines, etc. Communication technology is moving from 4G to 5G, but the growth rate of network bandwidth is being chased by data growth. IDC expects to have 40% of the Internet of Things by 2018 Data needs to be stored and analyzed at the edge, and 50% of the Internet of Things will face bandwidth problems. Facing the explosive growth of data volume and the bandwidth challenge, Intel realizes that more computing is needed. The edge is going.
Secondly, the proliferation of smart devices triggered by the Internet of Things has become a consensus. Almost all industries have imposed higher requirements on the speed of data processing. Various types of application scenarios cannot tolerate network delays and computational delays. In the case of autonomous driving, for example, if the camera recognizes that there is a pedestrian walking past the front of the car, the camera recognizes the image, then compresses the image, and then transmits it through the network to the data center for analysis and makes appropriate judgments. Far from the end, the judgment result needs to be transmitted back to the front-end vehicle through the network again, and corresponding braking operations are taken. If such a closed-loop is followed, when the car makes a final decision, it may have caused unavoidable losses.
In addition, data owners’ awareness of data privacy protection has been gradually improved. They do not want to upload data to the cloud and share it through third parties. They hope that these data will be processed locally. Therefore, some privacy protection requirements are relatively Higher application scenarios require data to be processed at the edge.
Edge Collaboration, Load Consolidation, Artificial Intelligence: Intel's Edge Computing Future Perspective
At present, many computing processes take place in the back-end data center. Intel believes that the application of the Internet of Things must require end-to-end capabilities. More and more applications will be pushed to the front-end for processing. Then, edge computing can be mitigated by The unique advantages of bandwidth pressure, timely response and protection of privacy play a vital role.
However, this does not mean that Intel believes that edge computing will replace cloud computing. The two will show complementary and coordinated development status: On the one hand, as the volume of data increases, the burden of cloud computing will increase. If you want to last, you need to The edge layer 'provides a helping hand' to do data preprocessing; on the other hand, in many application scenarios cloud computing can stand high and collect data to achieve comprehensive applications. For example, in the transportation industry, cameras can The captured trajectory of the vehicle is limited. Only one junction or area can be observed. If you want to fully check the trajectory of a car, it must be connected in the cloud through different cameras, and it is summarized into a more panoramic record.
In addition, Intel is also gaining insight into the other major developments in edge computing. With the increasing demand for edge computing from artificial intelligence, workload consolidation will become a general trend. Workload integration is the integration of small-scale edge computing into the central Servers, thereby reducing service costs and improving computational efficiency. Edge-ends can be aggregated into data nodes through load integration, and they are also control centers. It can be said that artificial intelligence also plays a role of benefit and back-feeding. On the one hand, artificial intelligence The development of a large amount of data is inseparable from the training of massive data. Many application scenarios of edge computing are a good endeavor for artificial intelligence. In turn, artificial intelligence can also give full play to its advantages, continuously tap data potential, release data value, and further promote the edge. Calculations move forward.
With the trend of data flooding, the rise of edge computing accelerates the process of Internet of Things. How to more effectively promote edge computing, Intel captures two major development directions of edge coordination and load integration. At the same time, Intel takes artificial intelligence to the edge to release data value. , also brings new opportunities for edge computing.