Distributed Operating Systems?

ayejay asks: "Are there any models/designs for a totally distributed operating system, possibly utilizing AI to learn patterns of use, resource need, and anything else that might be relevant? What -would- be relevant to such a thing? Given Napster and all the load balancing kernel enhancements and SETI@home type programs out there, it seems the idea is ready to be developed into a feasible paradigm. What do you think some of the major concerns/design issues are? I'm talking about nuts and bolts..." Now I'm all for distributed applications, but applying such paradigms to something as critical as the operating system seems to be taking the issue a bit too far. Would creating a 'distributed' operating system gain us any advantage over what we are currently familiar with?

Comments

Mosix

[1010011010]

Mosix is pretty cool, and will be even nicer when they have Distributed Shared Memory, Migratable Sockets and Direct Filesystem Access issues worked out (currently Mosix does i/o remotely through the home node, which makes it slower and loads the home node; DFSA allows remote nodes to access files locally rathen than via remote-I/O).

It provides preemptive process migration among cluster members. If you log into your "home node" and start a process, it will get migrated around the cluster according to its memory and CPU needs. Take a look at their remote monitor.

Currently it's Intel-only, but a mixed-architecture version would be sweet. Imagine a cluster of intel, alpha, PPC and sparc CPUs such that you log into any of them, run any Linux binary, and the loader cranks it up on the appropriate machien for you, transparently...

From the website:

MOSIX is a software package that enhance the Linux kernel with cluster computing capabilities. The enhanced kernel allows any size cluster of X86/Pentium based workstations and servers to work cooperatively as if part of a single system.

To run in a MOSIX cluster, there is no need to modify applications or to link with any library, or even to assign processes to different nodes. MOSIX does it automatically and transparently, like an execution in an SMP - just "fork and forget". For example, you can create many processes in your (login) node and let MOSIX assign these processes to other nodes. If you type "ps", then you will see all your processes, as if they run in your node.

The core of MOSIX are adaptive resource management algorithms that monitor and respond (on-line) to uneven work distribution among the nodes in order to improve the overall performance of all the processes. These algorithms use preemptive process migration to assign and reassign the processes among the nodes, to continuously take advantage of the best available resources. The MOSIX algorithms are geared for maximal performance, overhead-free scalability and ease-of-use.

Because MOSIX is implemented in the Linux kernel, its operations are completely transparent to the applications. It can be used to define different cluster types, even a cluster with different machine or LAN speeds, like our 100 processors cluster:

---- ----

Several Options...

[Christopher B. Brown]

• Mach was the "granddaddy" of distributed OS work, with most of the recent efforts going into GNU Hurd.
• There's Mosix that builds a NOW atop Linux
• The MIT Parallel and Distributed OS Group should be mentioned; efforts include the Exokernel
• Plan 9 has an interesting model for splitting work across "compute servers" and "file servers" and "display servers."
• Distributed Operating Systems lists lots of them...
• Sun's Spring was the basis for much of what is in CORBA;
• Sprite provided a Unix-like distributed OS that provided much of what is being used now to build journalling filesystems
• Amoeba was Tanembaum's successor to Minix; note that Python was one of the side-effects of the Amoeba project...

Each has some somewhat different insights to bring to the table; there is no unambiguous way of saying "this is all vastly superior."