Anunta tasked one of the research agencies to study the state of application performance management and monitoring in the Indian BFSI sector including banks, AMCs, insurance companies and brokerages. While some of the findings were what we expected, there were some interesting contrasts that the study threw up in terms of what these companies are saying and what the ground realities really are. As a first in a series that we hope to post on this survey, we’ll delve on some of BFSI’s top technology joys and sorrows.
WHAT THEY SAY: The survey found that given the rapid evolution and adoption of new technologies and applications for example, cloud computing, e-payments, mobile payments etc, respondents felt that challenges would only increase. This is a big pain point for 70% of the CTOs/IT heads that we spoke to.
BUT: It’s not something the sector can hide from. Take mobile payments for example – according to RBI figures, the volume and value of funds transferred through national electronic funds transfer (Neft) has been doubling almost every year. In 2010-11, the volume of funds transferred through Neft doubled to 13.23 crore and the value of transactions too doubled to Rs 9,39,149 crore.
Our take: All of the technologies cited above are ones that we believe will drive the future of technology infrastructures. Platforms like e-payments and mobile payments will mean added complexity to the application architecture which is already relatively under-managed at present. What this will mean is a higher level of integration and introducing new ways of monitoring enterprise and applications performance, all the while keeping a strict control on capex and costs.
WHAT THEY SAY: 76% of the respondents said they use automated tools to measure application performance.
BUT: 53% admit that there is a consensus between IT and end user measurement.
Our take: Measuring application performance is not enough. It needs to be measured from an end user perspective by translating technical SLAs into end-user SLAs and then enforcing these across the organization. However a dissonance between the IT organization and end-users on what a good measurement metric is, impedes the process.
WHAT THEY SAY: CTOs understand the importance of end user monitoring and 83% of the respondents measure the performance from end user side.
BUT: The metrics employed, capture end user experience broadly, but do not provide any detailed assessment of performance. They depict a reactive approach towards monitoring and are used mainly for incident reporting. Parameters include response time, application downtime, number of problem tickets and in a large number of cases, just end-user feedback.
Our take: While 83% seems like a healthy number, based on our experience in this sector, we also know that it is based on device level SLAs. There is an urgent need to become proactive in the approach towards end-user experience and issues monitoring. End-user SLAs when combined with technologies such as virtualization, allow SLA defaults and issues to be identified before they occur.
WHAT THEY SAY: It was observed that the loss in employee productivity was measured in terms of No. of volumes/ No. of people/ No. of hours lost due to incidents that cause a dip in application performance. Productivity losses are significant if the issue is unresolved in 30 minutes to 1 hour or when the network is unavailable for the entire day. In those cases, the productivity losses could go up to 30%.
BUT: Almost 56% of respondents agreed that they do not measure the business impact of lower application performance.
Our take: This is one of the biggest bugbears from an application performance management and measurement standpoint. Given that IT departments are often focused on keeping their TCOs low and ensuring system’s are running adequately (not perfectly), what they’ve lost sight of is the fact that IT needs to align to business. A drop in IT/ application uptime can mean significant revenue loss and brand erosion brought about by dissatisfied employees and customers. We’d addressed this issue in one of our earlier blogs but we expect it to remain an issue for a few more years.