Deploying GT.M based TCP services under an Internet super-server

Standards make the Internet work!

Internet super-servers listen at TCP ports for connection requests and launch the appropriate server process for each request. Using a standard way to deploy services separates the logic for implementing the service from system administration logic, such as limiting services to certain times, limiting them to certain IP addresses, protection against denial of service attacks, etc. Such partitioning of responsibility makes it easy to create and deploy services in MUMPS, since the logic only need to be concerned about the actual service. Using a super-server also also eliminates the need to have server processes waiting for each different service.

We will take a "deep dive" into deploying a service implemented in GT.M (http://fis-gtm.com) under xinetd (http://xinetd.org), a widely used Internet super-server with considerable functionality for security and logging.

This illustrates the "small is beautiful", conceptually compact design philosophy of GT.M. It is engineered to excel at what it does, with an open architecture that leverages the underlying operating system environment. Available as free / open source software (FOSS) under the GPL, GT.M on x86 GNU/Linux has few barriers to its adoption as a platform for MUMPS applications that scale up to the needs of the largest enterprises.