How Cloud and Edge Computing Can Collaborate

Since the birth of edge computing, there have been voices saying that edge computing is the ‘end’ of cloud computing. But after the verification of time, the relationship between cloud computing and edge computing is clearer: because edge computing solves the application problems of cloud computing applications in the edge resources, and becomes an important support for cloud computing in the future development, edge computing and cloud computing will inevitably merge with each other, and the ensuing is the ‘cloud edge synergy’. Edge computing is an effective complement to cloud computing Real-time or faster data processing and analysis. Data is processed closer to the data source rather than in an external data centre or cloud, thus reducing latency. For example, self-driving cars create a large amount of real-time data, much of which needs to be shared with neighbouring cars, uploaded to the cloud for computation, and then down to the end device, with unacceptable latency in data transmission. By utilising edge computing devices, it is possible to ensure that the information is processed quickly and responded to correctly, and that it is quickly passed on to other vehicles. Lower costs. Enterprises spend less on data management solutions for local devices than on cloud and data centre networks. Less network traffic. With the increase of connected devices, a large amount of real-time data will be generated, according to IDC forecasts, in 2020, the total amount of global data is more than 40ZB, a large amount of data to be uploaded to the cloud for calculation, the network transmission pressure will be more and more intense, while the process of edge computing, and cloud servers do not have much data exchanges, only a small amount of valid information needs to be uploaded to the cloud.

Edge computing is an effective complement to cloud computing

Real-time or faster data processing and analysis. Data is processed closer to the source rather than in an external data centre or in the cloud, thus reducing latency. For example, self-driving cars create a huge amount of real-time data, much of which needs to be shared with neighbouring cars, and the latency of data transmission is unacceptable when the data is uploaded to the cloud for computation and then downloaded to the end device. By utilising edge computing devices, it is possible to ensure that the information is processed quickly and responded to correctly, and that it is quickly passed on to other vehicles.

Lower costs. Enterprises spend less on data management solutions for local devices than on cloud and data centre networks.

Less network traffic. With the increase of connected devices, a large amount of real-time data will be generated, according to IDC forecasts, by 2020, the total amount of global data will be greater than 40ZB, a large amount of data to be uploaded to the cloud for calculation, the network transmission pressure will be more and more, while the edge of the process of computing, and the cloud servers do not have much data exchanges, only a small amount of valid information to be uploaded to the cloud, and therefore does not need to take up too much Therefore, it doesn’t need to take up too much network bandwidth.

More efficient application operation. With less lag, applications can run faster and more efficiently.

Runs offline and supports intermittent transfers. Reduced reliance on the cloud also means that some devices can stably operate offline in areas with weak signals or even no network service; when it’s time to upload data, simply move the device to an area where there is signal coverage to upload the data to the cloud. Scenarios such as an oil rig in the ocean or an aeroplane in the air are specific areas where there is a severe lack of network service.

Security and Compliance: In May 2018, the European Union passed the General Data Protection Regulation (GDPR), which has been described as the strictest data protection law in history. As data collection and computation are carried out locally, sensitive information can be transmitted to the cloud without going through the network, effectively avoiding data leakage during the transmission process, and part of the information will also receive appropriate protection if the cloud is attacked.

Edge computing should rely on the development of cloud computing

Internet of things in the device generates a large amount of data, data are uploaded to the cloud for processing, will cause huge pressure on the cloud, so separate processing. At this time, the edge computing distributed in each node will be responsible for their own range of data computing and storage work. For application scenarios, this is not enough.

Take automatic driving as an example, the future computing model is a combination of edge computing and cloud computing, the edge side of the automatic driving special chip will sense the sensor data and immediately processing, decision-making; at the same time, these data after processing, but also in the cloud convergence, big data analysis, model building and editing, while doing large-scale simulation, in-depth analysis and machine learning, and edge-side equipment for updating and upgrading. The edge-side equipment will be updated and upgraded to make the edge-side equipment smarter. Algorithm + chip + cloud computing, constitutes the three core pivots of the future of automatic driving.

It can be seen that the processing power of edge-side devices in big data processing, big data storage, application development, machine learning and artificial intelligence cannot be compared with that of the cloud. At the same time, application design, development, testing, deployment, management and other functions in the cloud are the key to developing edge applications.

Cloud computing cannot be replaced by edge computing, and the two complement and synergise with each other.

Combined with the above examples, it can be seen that the equipment providing edge computing capability is mainly in the front-end, responsible for real-time data collection, calculation and processing. However, most of the data is not disposable data, and the processed data needs to be retained in the system for algorithm training, data validation and other purposes. This requires a large-capacity ‘container’, which is not available in edge computing. In this ‘container’, the data will be stored for big data mining, algorithm training, user personalisation and so on, all of which are non-real-time requirements, and the data will be transmitted to the terminal equipment after the completion of these operations, thus further improving the quality of service. This ‘container’ is cloud computing, cloud computing to do big data analysis and mining, data sharing, while the algorithm model training and upgrading, upgraded algorithms pushed to the front end, so that the front-end equipment updates and upgrades to complete the closed loop of independent learning. At the same time, these data also have the need for backup, when the edge of the computing process in the event of an accident, the data stored in the cloud will not be lost.

From an overall perspective, edge computing can not replace cloud computing, but also can not be separated from cloud computing. In the future, cloud computing will form a complementary, synergistic relationship with edge computing, edge computing needs to work closely with cloud computing to better meet the needs of various application scenarios. Edge computing will be mainly responsible for those real-time, short-cycle data processing, responsible for local business real-time processing and execution, to provide high-value data for the cloud; cloud computing through big data analysis, responsible for non-real-time, long-cycle data processing, optimisation of the output of the business rules or models, down to the edge side, so that the edge of the edge computing to meet the needs of the local community, and at the same time to complete the application of the full life cycle management.