Research Manager-Enterprise & IPDS, IDC India
Billions of interconnected
The heart of this change would be the datacenter that would essentially need to get smarter than its current avatar and at a pace that needs to outshine the pace of IoT adoption. Smart datacenters would essentially comprise of three components to answer the challenges posed by IoT, namely, efficient and automated resource (both active and passive) management, smart architecture, and agile and secure infrastructure. While resource management would essentially comprise solutions consisting of intelligent PDUs, sensors, DCIM (datacenter infrastructure management) software, productivity, and efficiency solutions that are driven by data and insights are looped back in a continuous improvement framework. This would essentially mean the datacenters to be necessarily going through the phases of smart IT-ready, enabled, optimized, and continuous improvement stages to arrive at a stage where they are able to integrate extensive analytics in their operations and autonomously provision resources as per the demand. This would become important in the IoT age since with interconnected things increasing at an exponential rate, the infrastructure would need minimum human intervention but at the same time need to connect multiple management systems internally to be able to cater to the needs of that scale.
Another important aspect would be to broaden one's horizon and consider data center as a pool of resources rather than as a building housing the infrastructure. It could be physically at one or multiple locations and essentially could be located in geographically dispersed locations as well. IoT in essence would engage a lot of edge analytics and computing since data in motion and semi-motion would have to be constantly analyzed. This translates to workload-based architectures that can support edge analytics and network realignments to drive constant changes. This in turn would result in the growth of new virtual devices and functions (performing only functions and not present physically, e.g., virtual router), cloud-based hybrid architectures, enhanced storage pools, and edge computing. A mix-and-match is essential and inevitable as we look forward to embrace benefits from this interconnectivity.
Lastly, the agility and security of the resource pool (infrastructure) itself would be the most important aspect as we get interconnected. It is envisaged that millions of devices would enter and leave this interconnected chain every day in future for different functions and at different times and hence agility of the infrastructure would necessarily define the customer experience at the front-end. Seamless interconnectivity would also include aspects like middleware becoming all the more important to enable the layers above and below to communicate and understand each other. Security and privacy would essentially need to be seen in a new avatar for this infrastructure and resource pool to operate seamlessly without fear. While security would essentially comprise of concepts like layered approach (e.g., n layers comprising both physical or virtual or only virtual as the case may be), defense in depth, and behavioral pattern analysis, privacy would essentially need a new set of frameworks and architectures being enabled in the core (or at the planning stage) itself so that this interconnectivity becomes secure and is compliant to privacy as desired for a particular information, function, application, or service. With IoT bringing all these radical changes to our lives and infrastructure getting ready to face this battle, no one solution or approach is expected to dominate alone. It would be a partnership and alliance approach that would prepare the DCs for the