Why peripheral computing will play a key role in the enterprises
Medical devices, driverless cars, industrial robotics — number United with the Internet systems is increasing day by day and this growth is unlikely to slow down. New ecosystem provides obvious benefits to companies, however, the introduction of the “Internet of things” promises and difficulties. All connected devices are creating large amounts of data, so you first need to think carefully about issues of data management and to ensure appropriate investment.
According to the report Age Data prepared by the IDC under the sponsorship of Seagate, the amount of data worldwide will grow by 2025 to 175 zettabytes, with more than 80% of this volume falls to the share of commercial organizations. With network connectivity increasing numbers of devices increases moving data across the network, creating serious problems for certain systems (for example, self-driving cars and systems used for the provision of financial services), where the critical instant data transfer with minimal delay. Decision problems become peripheral computing by moving computing power from the core to the periphery of the data can be processed faster and more efficiently.
However, it is not so simple. Peripheral computing require new, more complex ways of organizing the it infrastructure, and to begin to enjoy the benefits, the promises of a new era of distributed networks, one should decide a number of important issues related to data management.
Exponential data growth and complexity
In the Gartner report “How to overcome the four main challenges associated with peripheral computing” says: “By 2022, more than half of data generated by businesses will be created out of data centers and clouds, but it will be other data types than those that are created in systems that today are associated with the processing and control”. In this regard, there are additional difficulties, because each company has their own needs and objectives work with the data and there is no single industry standard for the IoT ecosystem management covering the periphery and the core.
Many organizations have now moved at least part of their data in public cloud for the benefits it offers and the ability to quickly scale. However, with the accelerated growth in data use public cloud can lead to a range of issues relating to security, to increase the level of complexity and cost. Private and hybrid cloud can solve this issue but it should be borne in mind that their use causes difficulty in data management in the periphery and in the nucleus. Eventually there will be innovative software and hardware solutions to overcome these difficulties which will allow us to drop some of the stages of work with the data will facilitate and accelerate their transfer and will increase the reliability of the whole process.
It’s very simple: the more you network, the more opportunities to attack her. In any enterprise the advantages of peripheral computing will need to match with security risks and take measures to reduce these risks. If the data are in the public or private cloud, generally well protected, then the data in the periphery do not.
However, it can be noted that most companies today can use many solutions to improve data security. The key role here is played by the hardware: no matter how secure your software is, if whoever got in their disposal of physical storage media has the ability to retrieve them. Encrypting network data during storage should become an accepted standard, and encryption keys must be placed separately from the data. If the data is in storage will be encrypted, even in the case of compromise of the network itself, the information will remain safe. In the coming years, the hardware protection will be a critical component of an overall strategy to ensure data security in enterprises, and those companies in which think about it today, in the future will be in a better position.
The analysis of the value of available data
With increasing growth the amount of data stored on the periphery, enterprises will be difficult to obtain a complete overview of all available information and analyze its value. For example, the smart factory is a manufacturing enterprise can be a variety of devices, each of which generates data of its type. These data can be used to improve production efficiency or to detect failures in real time. The difficulty is to prevent the fragmentation of all generated data.
An additional difficulty is the fact that these different types have for enterprises of different value. According to Gartner, the average data bytes in the periphery will have less value than the byte in modern data centers. The reason is that a significant proportion of the data represents a “noise” to filter which requires pre-filtering or basic analyses. Consequently, it requires additional measures to assist enterprises in analyzing data in the field and their safe transfer over the network to the destination.
As corporate networks proliferate and become more complex, to evaluate the importance of data becomes increasingly difficult. Storage and data management in the periphery in the near future will play an important role in controlling the growth of data in enterprises, however, this will involve a separate class of difficulties. It is a critical moment: many companies are just entering the path of implementation of systems of peripheral computing. But if you are already thinking about how data management and security, as well as the analysis value in the long term, implement appropriate processes everywhere — from the periphery to the core — it will be much easier.
Information security, Internet of Things (IoT), Cloud computing, manufacturing
Journal: Journal IT-Expert, a magazine Subscription
About the authors
the head of the Russian representative office Seagate Article