Mihai Rotaru

Managing Director, Migratory Data Systems

Mihai is the project lead and developer of Migratory Push Server, a high performance, reliable, resilient real-time HTTP streaming server that can handle a huge number of concurrent users. It offers extreme vertical scalability supporting up to 1,000,000 concurrent users on a single entry-level server (benchmarks). In addition, the system scales horizontally with built-in load balancing to meet any growth in number of users.

Migratory Push Server is suitable for applications with lots of users to deal with poor scalability, high bandwidth consumption, and latency issues of the existing techniques used to achieve real-time data delivery. It has all the enterprise features needed for large deployments including load-balancing, fault-tolerance, guaranteed message delivery, and advanced monitoring.

Current Projects and Contributions

Migratory Push Server

Why Comet?

It’s an efficient method for real-time data delivery with zero installation on the client-side.

Sources of Inspiration

The need for extreme scalability.

Published Stuff

Here is a selected list of the articles I’ve published as a co-author during my Ph.D. research:

Molecular Interaction. Theoretical Computer Science, vol. 289(1), 801-827, Elsevier, 2002
A pi-calculus machine. Journal of Universal Computer Science vol.6 (1), Springer, 2000
Formalism for Distributed Systems with an Accurate Graphical Representation. In Proc.
4th SNPD’03, ACIS, 196-202, 2003
JC-Nets. In M.Margenstern, Y.Rogozhin (Eds.): Machines, Computations and Universality, LNCS 2055, 190-202, Springer Verlag, 2001
Ph.D. thesis: Semantics of the Asynchronous Concurrent Processes

Lifetime Accomplishments

Ph.D. in Computer Science.

Acquired substantial experience in the real-time financial software by having designed and developed software products used by tenths of major global capital markets institutions and investment banks.

Created Migratory Push Server as an extremely scalable Comet server able to push real-time data to 1 million concurrent users from a single instance running on entry-level hardware.