Peripheral computations: a friendly match of "fog" with "clouds"

    One of the main trends in the development of information technology in the past 20 years is the transfer of complex computing from a local computer to remote servers that are connected to it via computer networks. It all started with the concept of "network computer", which then turned into cloud computing.

    And after this logical technological evolution, we again hear the talk that it is better to transfer part of the calculations to local devices. This is the so-called peripheral computing, or Edge Computing. What is it - the further development of technology or a turn back?

    From classic data centers to Edge Computing

    At first, the concentration of computations in one place seemed like a great idea - of course, provided that the availability of the data center (data center, data center) was constantly increasing. And now the moment came when the fulfillment of this condition required a disproportionately large expenditure. The well-known Internet of Things (Internet of Things, IoT), which dramatically increased the amount of information generated and processed, added fuel to the fire. Plus, the need for digital technologies has increased - they are increasingly needed to manage operational processes, and the delayed response time of the cloud in the traditional approach turned out to be unacceptably large. Classic data centers and communication channels gradually cease to effectively cope with the information flows that are growing every day.

    Further centralization of computations requires expensive measures to increase equipment availability.

    As a result, it has become necessary to process some of the data closer to the producers and consumers of information. The scheme is as follows: only the data that is impossible or too expensive to process in a different way come to the large data center for processing. This is Edge Computing, or Peripheral Computing.

    Account in favor of Edge Computing

    This is like a football game. A team on the field is all kinds of devices connected to a computer network. She has a manager who solves strategic tasks, for example, is looking for the best places to train or organizes a transfer of "star" football players. This is a “cloud” or a large data center. There is also a team coach (Edge) - he communicates with the players and solves tactical issues related to the game. In the case of peripheral computing, its function is performed by the local micro data center.

    If the team is an amateur or is known only at the regional level, then the functions of the manager and the trainer can be combined by one person (cloud or local data center). But if we are talking about a professional team that plans to declare itself at the federal level, one cannot do without a division of duties between a coach and a manager. And if the sports club has several teams, each with its own coach? It is unlikely that the manager will be able to communicate with each team - he is busy with organizational issues of the entire club. The coach informs him about the situation in the team - he provides only strategic data important for decision-making. At the same time for the selection of such data from the tactical and their assessment coach uses his professional competence. This is exactly how micro data centers act.

    Peripheral computations can be implemented in three ways. The first and most popular is to create a micro-data center, which will become a kind of intermediary between users and a large data center. The second is to increase the “intelligence” and integration of IoT-connected devices. The third is to combine the first two. In this case, the computational function of the micro-data center will be partially performed by “intelligent” actuators.

    Schneider Electric EcoStruhture Ecosystem is based on IoT. Therefore, it is not surprising that for building a modern and reliable infrastructure, Edge Computing is meant to be deployed.

    Clouds are no longer needed?

    In general, peripheral computing does not imply the abandonment of the "cloud". The data that is processed by local devices may well be sent to the “cloud”, just the calculations will be carried out there to a smaller extent. The resulting hybrid IT environment provides more opportunities and flexibility in the organization of calculations and data storage, which undoubtedly has a positive effect on their cost, continuity and availability.

    With the use of technology Edge Computing load on the "cloud" is reduced.

    MicroCODs connected to the same “cloud” can communicate with each other directly. Thus, they are also included in the cloud structure of distributed computing and data storage. In order to designate such structures in the main part of the “cloud”, experts sometimes call the structure formed by them “fog”. At the same time, if one of the “fog” elements fails, the computational structure will not stop working completely, but simply slow down the speed of the calculations. This is the fundamental difference between peripheral computing and the long-used principle of parallelization of information processing, where the failure of a single computing node leads to a halt in the operation of the entire system.

    For example, if you use Edge Computing in production, then you can not be afraid that a temporary suspension of communication with the data center or the "cloud" will lead to a stop of the conveyor. The current tasks of managing production processes will be solved by micro-data center. Just for a while, the feed to the “cloud” or data center will stop, on the basis of which the tasks of optimizing production processes are solved, but the possibility of local processing, storage and data management will remain.

    What do we have

    MicroCODs - the most common and convenient way to organize peripheral computing. They are in a variety of "appearances" - from one rack to entire complexes, supplied as fully integrated engineering systems in container or Non-ISO type buildings. The most productive models of micro-data centers are not at all "micro" in scale, but for the needs of Edge Computing, such complexes are not often in demand. Anyway, solutions of high factory readiness on any scale, including microCOD, provide customers with additional benefits in the form of shortening the project implementation time, guaranteed results at the solution level, rather than components, simplifying the design, planning, and coordination of work on the site.

    Schneider Electric MicroCOD Series (from left to right) FX, CX and SX.

    Speaking of solutions, it is worthwhile to consider three series of micro-data centers: FX, CX and SX.

    The SX series makes it possible to create solutions of very different performance. However, such equipment can only be installed in a pre-prepared environment.

    MicroCODs of the CX series are designed for equipment with power consumption up to 3.5 kW. They look stylish, have several variants of decorative design, are equipped with an active ventilation system and air filters, have excellent sound insulation properties, which allows them to easily fit into any office interior.

    The FX line, or Smart Bunker, stands out. These micro-data centers are suitable for working in difficult conditions, as well as with increased security requirements. They have a multi-layer metal case with built-in thermo-insulating gaskets, and massive doors and panels with a multi-point locking system with the possibility of remote dispatching and video surveillance reliably protect against unauthorized penetration inside. In addition, a full-fledged cooling system is provided using various options for its power and type, an uninterruptible power supply system in several power configurations and autonomy times, a power distribution system with multi-stage and intelligent protection functions, and an environment monitoring system. And finally, there are backup options for any components of the integrated engineering infrastructure. Generally, the FX series in hardware is reminiscent of a full-fledged, high-tech, secure, and fault-tolerant data center, but on a micro-scale. It can easily be placed in unprepared premises, for example, in the factory floor, in the basement or in the attic of the building.

    However, sometimes there is not enough time to deploy even micro-infrastructure. For such cases, Schneider Electric has developed a special line - Micro Data Center Express. These are ready-made sets of engineering infrastructure of increased fault tolerance in the SX series SP version cases, which allow to create micro-data centers suitable for transportation with IT equipment installed inside. Moreover, the admissibility of such transportation is confirmed by Cisco. That is, the delivery to the object of a fully finished data center in the constructions of the SP version allows you to keep the company warranty from the manufacturer of the equipment.

    You can quickly transfer computing power from one geographic point to another if changes in your business require it. But more importantly, you can calmly assemble and carefully adjust the IT complex of the micro-data center on one site, and then bring it to the place where it will be used. For example, from a central office to a remote branch. At the same time, the solution will be immediately ready for operation; all that remains is to connect it to the power supply and data transmission network. To install the device, it is not necessary to stop production processes for a long time or put them at risk, send specialized engineers to start and set up the complex, and so on.

    For the selection of the basic solution or the configuration of an individual product, by choosing the parameters in the Local Edge Configurator section, we recommend using . In addition, you can always go directly to the Russian office of Schneider Electric, get advice, formulate a task in any form convenient for you and get a solution.


    The development of peripheral computing does not mean that the IT industry seeks to completely abandon the "clouds". On the contrary, it is thanks to this technology that cloud computing will come to areas where they have not been used before, but in a different format. For large enterprises, the sharing of “cloud” and peripheral computing will allow to get rid of a number of problems and disadvantages of cloud technologies, to significantly optimize the amount of transmitted data, their processing speed, reliability and availability.

    Also popular now: