For businesses to be agile and respond quickly to changing market conditions, they need to provide business users with real-time and near-time operational data. That means harnessing data from devices and tackling the latency challenge. In 2020, we will see more organisations shift their design thinking from services and systems for people to services and systems for machines. The move to machine-to-machine (M2M) systems also means processing is moving to the network edge, where the data is.
Action at the edge
Organisations are experimenting with extending data clusters to the edge to reduce latency, gain operational efficiencies and improve products and services. Fast food company Chick-fil-A is running Kubernetes on 6,000 kitchen devices in all 2,000 of its restaurants. This is part of the chain’s internet of things (IoT) strategy to collect and analyze more data to improve throughput, operational efficiency and, most of all, customer service.
The move to M2M dovetails with the notion of catering to local markets using local data. This will foster new design architectures that take into account privacy, security and regulatory concerns regarding the “frame of reference” of data — i.e., the notion of localised data being more valuable than global data in terms of its usability and mining. When organisations consider modernizing their IT operations and changing the way microservices are deployed — especially large multinationals with a wide global reach — it becomes critical for them to consider the advantages of targeting a geography locally, and realizing that there are new tools and ways of doing this.
Another facet driving the change in design thinking is the need to make maximum use of computing resources. Whereas the decision frequency of a person ranges between 1-15 hertz, which means that people can decipher information and make a decision in about ½-1 second, today’s microprocessors can operate at gigahertz and process information in nanoseconds. If these processors are not operating as fast as they can, they are just space eaters. Organisations want to keep their processors as busy as possible, which means designing for billions of decisions or operations per second. Otherwise, they may end up paying for unused capacity.
Signs of M2M
Two examples of M2M architecture are SAP Leonardo and KubeEdge. Through SAP Leonardo intelligent technologies and capabilities, SAP is integrating its ERP applications with IoT platforms, combining traditional IT services with M2M capabilities. Importantly, SAP can address broader markets than niche IoT platforms.
KubeEdge, built on Kubernetes, is an open source platform for building edge computing solutions that extend to the cloud. The platform supports network, application deployment and metadata synchronization between the cloud and edge. It extends the Kubernetes ecosystem from cloud to edge and provides benefits such as lower latency, low resource consumption and applications at the edge that can run in offline mode.
The road ahead
We are starting to observe a shift in IT design from IT for humans to IT for machines. These design patterns deliver richer experiences because they enable substantially more processing in the same experience time. Design shifts will lead to changes in batch processing and stream processing architectures, which are constantly being updated and reimagined with better M2M capabilities. Data and analytics will continue moving to the edge where the machines are, to analyze the massive influx of IoT data and provide maximum throughput with minimum latency. Rapid deployments of these transformational architectures may not be immediate, but over time these new architectures will be a forcing function for IT modernization.