Ksplice provides updates without reboots [LWN.net]

Ksplice provides updates without reboots

Posted Jul 10, 2009 16:41 UTC (Fri) by jbarnold (subscriber, #51843)
Parent article: Ksplice provides updates without reboots

As I mentioned on the LKML last year [1], the Microsoft "patent" that was pointed out is actually a rejected patent application, not a patent. It received a "final rejection" from the U.S. patent office in 2006.

You can browse the relevant documents at http://portal.uspto.gov/external/portal/pair (The application number is 10/307,902).

[1] http://lkml.org/lkml/2008/4/29/708

(Log in to post comments)

Ksplice provides updates without reboots

Posted Jul 13, 2009 1:34 UTC (Mon) by jhhaller (subscriber, #56103) [Link]

[I posted a similar comment on Steven Vaughan-Nichols' Computerworld ksplice article]

While the patent office has rejected Microsoft's patent, Microsoft is appealing it to court, so it's not over yet. Also, the reason for the rejection was a prior HP patent application. Presumably, HP could object to ksplice, although they are friendlier to Open Source, and their patent was related to dynamically patching code related to missing hardware instructions, not patching code. But, I think there are many other patching frameworks which are likely to be prior art. IEEE Software, March, 1993 had a survey of patching frameworks dating back to the 1970s.

In my opinion, the main innovation of ksplice isn't the dynamic patching, but the automated discovery of what and where to patch, as well as allowing changes to the number of function arguments. However, until shared libraries and long-lived processes are addressed, I think the business case for a service will be limited.