Published Date 11/8/18 8:00 AM
“Data Center World” (DCW) Singapore is a specialist event for the data centre industry; with many leading banking and finance businesses operating in the area, the attention has always been focused specifically on these areas, as indeed ICT infrastructure represents the essential backbone for these services. It is an important hub for the region, yet also a strategic event globally, as confirmed by the presence of many international visitors.
As already highlighted in a previous post, the volume of data generated by devices connected to the Internet, driven by IoT, is growing dramatically; this is forecast to double every two years, until reaching, according to some expectations, 44 zettabytes (just to give an idea for those who are less familiar with the topic, one ZB is a billion terabytes), with the number of connected devices reaching 80 billion in 2025. This rapid acceleration is due in particular to developments in self-driving cars, which alone could generate 4 TB of data every day.
The Internet, as already mentioned in the previous article, is becoming a bottleneck, and the performance required in terms of processing response is also increasingly demanding, with latency times that need to be reduced (again a good example is the management of self-driving cars). For example, if you could send a bit from San Francisco to New York City and back on a fibre-optic cable without any routers or repeaters along the way (something that is not realistic), it would take 42 milliseconds (latency); this is just one bit without any processing, so it’s easy to imagine how long it would take with complex processing operations in between.
Hence the need is clear for the development of “Edge Data Centres”, which provide local processing and a rapid response, and can therefore be an optimised communication node with the “cloud”. This change in approach is seen as the reverse of cloud computing, or “taking processing to the data”.
Many presentations at the Singapore exhibition featured several areas of interest for the “edge”, from industrial production (data centres as the “digital twin” of the machine tool) to retail (virtual reality for purchases), transport and smart cities, household/commercial entertainment systems (gaming), and even more critical areas such as military applications; this could intersect with the spread of 5G networks and see an enhancement of telephone infrastructure installations with the addition of edge computing.
It certainly does not mean a disappearance of cloud computing, but rather the latter completes the concept, providing coordination, storage, backup, machine learning... To make an illustrative comparison with computer architecture, the cloud represents the hard drive, while the edge is the RAM.
The need to develop more complex and “immersive” architecture that blends together both the network and processing is pushing industry to develop specific solutions. At DCW Singapore, many important players in the data centre cooling industry presented solutions for micro data centres, i.e. compact solutions, including containment infrastructure (usually a rack), the power supply and UPS, cooling, access control and fire prevention. The rack housing can be suitable for installing different types of servers with very high densities; the solution can be used usable in different contexts, from production plants to the telephone shelters and commercial installations, without the need to build a suitable room from scratch. In line with this are the so-called mini data centres that represent modules with multiple racks, multiple air-conditioning units and power distribution systems, completed by a system for containing the hot/cold aisle and safety and fire management. These modules are also suitable for the development of larger data centres by simply combining several of these in a centralised installation.
Example of a mini data centre
Visit CAREL IOT website data centre page.
DCW London - Continuous evolution of Data Centres: IoT, edge and “fog computing”
DCW Singapore: opportunities for Data Centre energy savings in tropical regions
From the Internet of Things to Intelligence of Things