From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing (Rated Start-class)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 ???  This article has not yet received a rating on the project's importance scale.

Posted in article by So what now with this exokernel separation of protection from management? How about runaway processes/apps?

The kernel doesn't decide how the allocated resources are used. It only allocates them and guarantees the integrity of the allocation. Runaway process, as noted in the article, are penalized in the future or aborted. Gazpacho 08:31, 30 November 2005 (UTC)
So did you try to find out how it handles it before commenting? There are plenty of papers if you follow the Papers link on the MIT Exokernel page. In any case, the statement is correct, see

Graphical overview[edit]

The graphical overview should show the libraries inside the applications because they use the same address space.


The MIT exokernels manages hardware resources as follows:

1. exokernel manages
2. exokernels manage

I decided for the first solution, but it also might be the other way around


What is this type of kernels actually good for? I understand its for performance control mainly (as in low-power systems)? The disadvantages seem to be quite immense - user having to deal with the hardware, low portability (of programs), unless one builds a system of user libraries on top that provide the abstractions. But then whats the point? On the other side the exokernel must provide some level of abstraction anyway (e.g. block access to the hdd, instead of sending IDE commands) so what actually makes it an exokernel - for sure not the lack of in-kernel file system. User:misiu_mp —Preceding undated comment was added at 00:33, 27 August 2008 (UTC)

I believe they're running incredibly fast. On the other hand, they're only research kernels trying one end of the spectrum in a line of more and less controlling kernels. I'm not even sure the practicallity of Exokernels is proved. Rursus dixit. (mbork3!) 08:41, 18 February 2010 (UTC)
Isn't the original paravirtual Xen sort of a exokernel, especially in the services it provides a DomU? It was at least inspired by the idea, the heritage is the MIT project -> Nemesis -> Xen. Hga (talk) 11:42, 18 February 2010 (UTC)
Don't the research papers make the rationale pretty clear? Exokernels offer tremendous scope for optimization, and is as flexible as possible. In a sense, it's a superset of monolithic and microkernels. --Gwern (contribs) 20:58 18 February 2010 (GMT)
The MIT Exokernel code base was the starting point for the XOK/ExOS embedded system at the core of the Vividon/StarBAK Streaming Delivery Accelerator (SDA) appliance. We cleaned up the code, added support for specific hardware, added a mission-specific file system, customized the address map, and simplified the network frame handling. The result was carrier-grade reliability and outstanding performance, with a user-space code interface that was mostly OpenBSD compatible. Very usable, thank you. Dave Tuttle (talk) 21:02, 20 August 2011 (UTC)

Image cleanup[edit]

The image needs to be improved. It does not illustrate anything specific to exokernels; it could apply just as well to a monolithic kernel. It is also notionally upside-down, putting "low-level" code at the top of the diagram. -- Beland 20:29, 21 March 2007 (UTC)

Semi-tidied this image and moved it to Switched the Software and Kernel sections around as you suggested. Someone with more knowledge on the subject needs to edit the diagram to show what features are specific to Exokernels. bobbo (talk) 20:51, 16 July 2008 (UTC)