Some special applications
  • Some cool stuff here!

    We've done some really cool projects, because we were lucky enough to have clients who trusted us enough to take the risk. Many of these projects are unlike anything else which is available anywhere. Some are similar to existing systems but perhaps at one-tenth or one-twentieth the investment. All of them do what they intended to do excellently, and most of them are in use at their respective client sites round the year.

  • PRISM is an abbreviation for "Parallel Risk Machine". PRISM is a cluster of Linux computers which act as compute nodes in an HPC cluster, with a RISC Unix system as the hub. The goal was to deliver VaR (Value-at-Risk) computations for one of the world's busiest stock exchanges if trades closed at a rate of 1,000 trades/sec. This system is highly scalable because the load is highly parallelisable and compute nodes can be added at will. It is fault tolerant in the presence of compute node failure, and takes less than 50 msec to recover from a downed compute node. All nodes communicate with each other over an MPI layer. When PRISM was commissioned in 1999-2000, it was the only system in the world capable of performing real VaR calculations at those speeds in use in any financial institution or trading exchange. Read more

    Assembly line optimisation This project was quite unique for us -- our software interfaced with industrial electronics hardware to offer a total solution which operates reliably on the shop floor. Our system helps line supervisors and production managers to optimise work on the assembly lines in real time. In this industry, manufacturing productivity is measured in GTD (Garments per Tailor Day). A figure of 12 GTD is very high for Indian factories and 17-18 is excellent in the international context. Our assembly line optimisation system, together with our customer's efforts in fine-tuning processes around it, delivered an improvement of 1-1.5 GTD in productivity. Read more

    Web activity forensics One customer, who are one of India's most important financial institutions, operates many important Internet-facing applications for various statutory and government departments. We built what we can describe as a "CCTV at the gate" for their Internet-facing traffic. Our system watches and records all HTTP requests and responses, and then analyses the data to identify suspicious behaviour or unusual changes in access patterns. What is exceptional about this system is that it can report details per username, because it can map accesses to the username being used by users to log into the systems. And it can do all this without asking application developers to make a single line of code change in their applications. Read more

  • When we look at these projects, we know Merce is not just another software services team. We have what it takes to take risks and bring about real change. We have done projects which most other general-purpose software services companies would not have dared to touch -- yet, we consider ourselves a general-purpose software services team.

    Do you have it in you to work with us to make a difference in this way?

    See all case studies


  • Models of engagement

    We engage with customers in various models, depending on their requirements.

  • Case studies

    A collection of our case studies, indicating the range and depth of work we have done