What's your role at Sumerian?
I work as a Senior Analyst within our Data Scientist team. I work on a diverse range of analytical projects for our client base, helping them to gain business value out of their data.
What industry sectors are you working with?
We provide our services to a range of industries, though a lot of the work I do is with investment banks. This is a highly competitive market and gaining the edge with optimised IT is paramount - performance needs to be high and latency needs to be low. The economic crisis has meant that many companies, even the banks, cannot simply throw money at their systems - they need to know where to target investment precisely for the best gains.
What type of analytics have you been working on recently?
Recently, I've been doing a lot of work around virtualisation and private cloud. We're finding that more and more of our clients are coming to us with these challenges. Although virtualisation and cloud makes it easy to administer resources, it can be notoriously difficult to manage capacity effectively. However, if you can get a strong handle on it, the rewards can be immense.
For instance, by using virtualisation, batch processing can take advantage of the same resources that were used for intra-day processing capacity at a time when it is not needed, resulting in you making better use of your available resources and re-allocating the previously used batch resource somewhere else, or even decommissioning it. The added benefits are twofold: you get better utilisation rates from your IT, and at the same time you reduce your energy costs by using fewer datacentre resources. With the trend toward private and hybrid clouds (which are hinged on virtualisation) gaining momentum, this is a strategy that's likely to become increasingly commonplace.
What does Sumerian offer around virtualisation challenges?
The challenge we're seeing is that many of our clients are incredibly stretched purely managing their IT environments and business change. What Sumerian brings to the table is the ability to take vast quantities of data from these complex virtualised environments and distil it down to real, actionable insight.
Can you explain an instance where this is being used?
One of our clients is undergoing a large-scale change programme to overhaul their e-commerce channels. A key challenge for them is ensuring they have enough capacity for their planned business growth - and effective use of virtualisation forms a major part of this. One of the reasons they approached us was that the visibility they had into their virtualised environment and resource pools was not good enough to manage this well.
To explain further, a virtualised environment is split into pools with each one given a set number of resources that it shares out between a number of virtual servers, when a server needs additional resources it's able to "borrow" them from the pool and return them when the task is completed. In the work that we are carrying out for our client we have been able to show how they could re-configure their existing resources in a way that would give them greater capacity for growth and reduce the need for further expenditure.
Using our IQ service model, we capture volumes of web traffic across a number of channels and relate that to the load being placed on the virtualised e-commerce infrastructure. We are then able to see if some servers are over resourced for the workload they have been allocated, and can therefore recommend where resources be freed up or re-allocated to other servers in a way that provides greater available capacity.
How were the results of the analysis presented?
One of Sumerian's key strengths is the way that we can convey our results in a clear, visual format. For example, the way we show our clients how their virtualised environments are being used is by way of a "Virtualised Resource Allocation Chart" (see right).
The chart shows the relationship between maximum resource capacity (CPUs in Pool), entitled capacity (CPU Allocated), and actual usage (CPU Used). Our clients have told us that this visualisation allows them to see at a glance how to allocate resources to meet demand.
Tell us about your background and interests?
I commenced my career at Sumerian after completing a Masters in Operational Research from Strathclyde University and BSc in Mathematics from Warwick. One of my professional interests is the application of statistical analysis techniques to solve complex problems, for gaining a deeper, clearer understanding of the factors at play. The use of analytics is taking root in more and more areas of work and everyday life. It's not only used by businesses in areas such as risk analytics for investment banks and business intelligence for operational targets, it's also being used in areas such as supermarket promotions, improving patient care, and even sport.
The use of analytics in sport is a fascinating subject alone - it's been used in baseball for batting averages for decades, in Formula 1 to optimise racing strategies in real time - and a major growth market right now is in football. There's a lot of scepticism about analytics in football from people who believe that it's passion rather than tactics that win games, but the use of services like Prozone and Opta along with more amateur interests like Michael Cox's Zonal Marking are dispelling these myths.
You can take a look at these sites here to get an insight into how analytics is being used in sport: