Get a demo
I hereby consent to the processing of my personal data specified herein by CROC, for the purposes and within the scope set forth by the Personal Data Protection legislation of the Russian Federation, in conjunction with the activities performed and for an indefinite term.
Preferred communication method
Get a quote
I hereby consent to the processing of my personal data specified herein by CROC, for the purposes and within the scope set forth by the Personal Data Protection legislation of the Russian Federation, in conjunction with the activities performed and for an indefinite term.
Preferred communication method
Try for free
I hereby consent to the processing of my personal data specified herein by CROC, for the purposes and within the scope set forth by the Personal Data Protection legislation of the Russian Federation, in conjunction with the activities performed and for an indefinite term.
Preferred communication method
cloud.croc.ru
Learn from experts

Does every Cloud has a Silver Lining?

4 minutes 116

Andrey Tishchenko

Technologies may evolve for decades before either gaining momentum or achieving recognition of their potential among the general public. And when this happens, newly coined marketing terms are introduced to persuade us that they describe something truly breakthrough. Edge computing and Fog computing are in the middle of a similar situation.

A little while ago, building decentralized infrastructures was a common practice among large holdings – banks and large retailers, for example – to manage branches and accelerate transaction analysis and operation processing. Now, companies follow another trend towards data consolidation within one data center or cloud to simplify resource management. Meanwhile, edge and fog computing earlier considered a mere architecture of an application, now, are believed to be able to drive out clouds. Particularly, Gartner experts take this attitude, when analyzing IoT market growth.

Mess of definitions
To debunk the myth of dispersing clouds, we need a clear definition of terms first. The terms 'edge computing' and 'fog computing' are often used interchangeably, but edge computing is actually a component of fog computing. Edge computing occurs directly on a sensor or data collecting device. Fog computing is carried out between endpoint devices and data centers, i.e. within a branch's on-premise infrastructure or company's geographically distributed office.

Edge computing and IoT: when centralization is not required
Analysts got right the correlation between fog/edge computing and IoT development. Indeed, a market demand for processing tons of data coming from an army of devices raised a question of where to store and process it in a cost-efficient way. Data centers are not always the perfect candidates for the job. Stuffing expensive data centers with trifle short-lived information seems like overkill.

Let's take a manufacturing enterprise for example. It has systems installed to monitor temperature modes at blast furnaces. Data is updated daily using thousands or tens of thousands monitoring devices. The solution enables end product quality assurance in real time and ensures compliance with occupational health and safety requirements. Does it make sense sending such data to a head office across the Russia? Of course, it doesn't, since the data are valuable for local management.

Let's take another situation, where a company collects terabytes of video materials, for example, source data from video analytics system of transport organizations and corporate customer security services. It would be too costly to transmit such telemetry or surveillance data non-stop, as it would simply choke the bandwidth. Again, such transmission is as meaningless as the one from sensors at manufacturing facilities. Data analysis at locations where video recorders are installed will be sufficient, while recorded incidents can be cut out of the video stream. In other cases, transmitting the event log to the center instead of the entire video stream will be enough.

Edge computing and IoT: when centralization is required
The Internet of things can indeed drive the development of edge computing, but it doesn't mean it will replace data centers and clouds. On the contrary, the spread of IoT and some other technologies will gradually slow down without clouds. For example, artificial intelligence (another concept with a decade-long history, like edge/fog computing) can be developed on the basis of data from the above mentioned manufacturing sensors. If a company's goal is not limited to constant monitoring, but also includes discovering patterns in the operation of manufacturing facilities, then the company needs to periodically send data from devices to a cloud-based analytical system. This is the only way to ensure that the system will be capable of self-learning.

Another IoT use case that doesn't imply the death of a cloud is to constantly collect devices' operation data. For example, a cloud telemetry system covering tens of thousands vending machines across Russia. An analytical system aggregates data from machines and displays it in operator's mobile app. The operator always has a smartphone at hand to monitor the quality of product range, expiration dates, profitability of each sales point, and plan the most efficient delivery route. Another reason for the centralized management of this IoT product is that the analytical system has to receive fiscal data in real time as well.

IoT is a massive phenomenon capable of influencing the market of data storage and processing devices. Indeed, in some cases, data will be processed outside centralized data centers, but most likely, there will be a compromise in the form of multi-component structure based on various technologies, including cloud computing.

Don't miss the most important, interesting and helpful posts of the week

Success

scrollup