Data centre modernisation has reduced DC space usage by 30 percent for Deepak Fertilisers
Deepak Fertilisers has also been able to achieve a power usage effectiveness of 1.2, says Deepak Keni, Executive Vice President - Special Projects & Enabler, Deepak Fertilisers And Petrochemicals Corporation
A brief on certain achievements in the area of DCs by Deepak Fertilisers, for e.g. reduction in power consumption; achievement of SDG goals for DCs, etc.
Our main data center is based in the corporate centre in Pune and there are small data centers at the plant locations. The adoption of emerging technologies and innovation has been done at the main data center. We have been able to reach a PUE of 1.5 (The power usage effectiveness (PUE) is the metric to measure the energy consumption effectiveness of the data center), very near to the ideal of 1.2. In the last couple of years, we have made investments in modernising the data center in all aspects. The core infrastructure like the servers, have been virtualised and as a result the overall space of the data center i.e the real estate occupancy of the DC has been reduced by 30 percent.
Our data center requirements will only increase from hereon as we aim to gather more data from farmers, to provide precision agriculture solutions.
Challenges on the way in making the data centre truly software defined ?
Deepak Fertilisers has upgraded its data center facilities in the last two years and we are on the path of becoming truly a software defined data center (SDDC), in all aspects – network, compute, storage, etc.
The concept of a SDDC was always there, however with IoT and other technologies giving rise to more data generation is creating opportunities for enterprises to get insights from the same level of investments into the data center, which is an expensive proposition. But with making it software defined, the same data center resources are able to generate 5X, 10X and even more amount of returns from the insights they provide. Also compared to before, all layers of the IT infrastructure is now getting virtualised. Some years back, it was only the servers but now, the storage, network and other aspects are also getting virtualised to provide more capabilities.
Every company is at its own level of maturity in the adoption of these virtualisation opportunities based on the overall business strategy. Deepak Fertilisers is at the infancy of its data journey. We provide crop protection fertilisers and the current strategy was to focus on the dealers and the distributors and what happens after that, how it gets to the farmers is of less relevance to us. However, now the strategy is moving towards providing precision agriculture advisory services to the farmers. There are approximately 100 mn farmers in India and to provide them with fertilisers of the right nutrition at the right time requires a well rounded understanding of weather, crops, irrigation and a massive amount of other data points.
Emergence of Edge DCs in the times of the pandemic, when WFH and WFA has become a trend ?
Edge devices and edge computing is fuelling the demand for edge data centers. The humongous rise in the micro processor capacity of the mobile devices has empowered the applications in them to process data points many times more than a few years ago and thus to support this processing, the edge data center has to be placed at a nearby location to avoid latency issues.
The key to the success of edge data centers will be the ubiquitous availability of high network quality. The issue would not be much in Tier1 and 2 cities but in the interior areas, the remotest part of the country.
We as a company have five operating locations and there are small data centers in each of these locations with a focus on data center operations to make sure the manufacturing sites are up and running. The focus going forward will be on farmers and how can we provide them with the right fit based on their crop planning and weather.
Please discuss some use cases of AI in DCs
As we progress towards the journey of data centers getting software enabled, the best use case is how RPA can be used for automating hardware and software configurations; With software enablement comes massive data generation and again AI and ML can get more insights from the data in making operations more efficient depending upon the need of the organisation. It can be about the uptime requirement, getting alerts as a threshold is breached, when the cooling equipment is about to fail, etc.
Cyber security is a big use case. With active monitoring of the cyber security scenario, there are certain characteristics that come out and precede a cyber security breach. The algorithms feeded and designed with those characteristics can help in predicting an impending cyber security breach attempt on any particular organisation.