Using edge data centres to create decentralised IT resources
The volumes of data that now need to be processed are soaring as a consequence of digital transformation, so companies need a quick and easy solution for establishing new data centres directly, where that data is generated.
By Clive Partridge, Rittal technical manager for IT Infrastructure
Modular edge data centres offer the ideal solution here. The example of retail shows how the point of sale can be optimised using data-assisted processes and why this calls for decentralised IT resources.
Edge data centres are decentralised IT systems that deliver computing power directly to the location where the data is generated. They are situated in the immediate vicinity of the data sources – which helps ensure exceptionally fast initial data processing – and are also linked to cloud data centres for downstream processing.
Software applications in connected data centres ultimately use this up-to-the-minute data to perform analyses that require high levels of computing power.
Data is becoming increasingly important in physical retail spaces
Additional computing power enables companies to evaluate data relating to customer behaviour and enterprise resource planning more quickly and precisely. For example, a retailer can compare the sales in its nationwide branches using surveys from social media platforms in order to identify new trends. Alternatively, on entering a shop, customers – provided they have consented beforehand – can be identified via their smartphone and greeted with offers that are tailored specifically to them. This, too, requires an IT system that responds in real time and can access large volumes of data.
In general terms, edge data centres help companies to evaluate all customer data and thus optimise sales. They intelligently network branches that are spread out geographically with regional warehouses and a central data centre so as to optimise product availability at the point of sale (POS). Retailers can thus harness networked edge computing to increase the availability of products, optimise logistics and use customer preferences to regularly improve product displays at the POS, for example. The continuous and rapid availability of data gathered via edge computing makes it possible to manage customer behaviour more effectively. If necessary, this can be done as often as every day based on up-to-date data.
Retailers can also use the additional computing power and real-time stock tracking to optimise their supply chain management. In this case, long-term analyses help identify patterns in sales and thus provide plenty of notice regarding when specific products may be affected by bottlenecks. Without predictive analyses of this kind, there is the risk of losing customers, who switch to the competition because they can’t get what they want.
Networked IT infrastructure at the POS
To track goods and customers, retailers are installing networked sensors or using cameras to analyse patterns of movement. This is creating an Internet of Things that utilises a large number of sensors and data sources to generate a continuous data stream. Retail chains use sensors, for example, to identify the positions on the shelf where each product sells best. This also involves developing the supply chain to the extent that a shop reorders new products in an automated process. In the future, a growing number of companies will be using edge data centres to expand the requisite IT infrastructure at the POS. According to market analysts from IDC, edge IT systems could be processing and analysing 40% of data from the Internet of Things throughout industry by 2019.
What types of edge data centres are available?
An edge data centre is designed so that companies can adapt it to the required performance level using preconfigured, standardised modules. Climate control and power supply modules, stable IT racks and robust security components are already aligned with each other – this is particularly important for sites that do not have a specific security concept at building level – i.e. access controls or airlocks, for example.
If factors such as dust, humidity or dirt also pose problems at the site – because industrial production is carried out there, for instance – then the IT racks should have a high protection category, such as IP 55.
Edge systems come in a wide range of output classes depending on the requirements and area of application. Edge gateway systems, for example, consolidate data directly on site and then initiate its transfer to downstream cloud data centres.
However, initial evaluations can also be carried out close to the data source. For instance, smaller systems for retail can perform tasks such as the initial aggregation of sensor data in a department store, supermarket or shopping centre, while powerful edge data centres can also be utilised that significantly increase the computing power at the relevant location. The latter may be necessary if retailers want to offer their customers elaborate product presentations based on virtual and augmented reality.
The technology used in these edge designs can vary greatly – from a basic service rack to a specially secured IT rack with an additional protective cover. If more power is required, a high-performance edge data centre based on a modular data centre container with weather-resistant and fire-resistant covering is the answer. The solution is then installed in the immediate vicinity of the location where the data is generated, either inside or outside buildings. With appropriate cooling technology, it will support an output of up to 35kW per IT rack.
Thanks to their steel walls, IT containers are both stable and secure. Their excellent mobility also makes them highly flexible and means powerful data centres can be installed anywhere on company grounds or inside warehouses.
Requirements determine the configuration
If edge systems are being used to boost on-site computing power, the first step is to specify the associated business objectives. Technical and IT experts use this information to define the necessary software applications and it’s then possible to determine the configuration of an edge data centre based on this list of requirements. A number of criteria need to be taken into account during this process, for example, edge systems must be quick and easy to use in order to meet technical requirements promptly. The ideal scenario is for the manufacturer to supply a turnkey, ready-assembled system, complete with cooling technology, for plug-and-play connection to the power supply and network technology.
Edge system operation should also be automated and largely maintenance-free to minimise running costs. This requires comprehensive monitoring that covers the power supply, cooling, fire detection and extinguishing. The necessary protection category is determined by factors such as location and how fail-safe the system needs to be. It is also important to use a monitoring system that covers enclosure/rack doors as well as side panels; electronic door locks have the added benefit of making it easier to ascertain which staff had access to the IT and when.
During remote maintenance or emergencies, it may be necessary to completely power down the system, which means having to interrupt the power supply. Switchable PDUs (power distribution units) are required for this purpose.
Enhanced security with edge
Edge data centres can be installed in a room-in-room environment for the toughest security demands and a security room of this kind offers maximum protection in the event of fires or highly contaminated surroundings. Outdoors, it should also be ensured that the protection category supports reliable IT operation across a wide range of temperatures, for example from -20 to +45°C.
Suppliers such as Rittal have developed a modular concept for these varying requirements and companies can use a modular system to create the ideal solution for their needs.
Rittal adopts a holistic approach when seeking a solution, working with partners such as ABB, HPE, IBM and the German cloud provider iNNOVO so that customers get all the services they need from a single source. The resulting pre-defined, standardised all-in-one edge system can be augmented with active IT components and ‘as-a-Service’ options in a turnkey solution. The retail sector is therefore able to use continuously updated data to optimise the customer journey, and thus secure customer loyalty on a long-term basis.