Pre-Grant Publication Number: 20070288718
Collaborate on the process of community review for this application. Posting will not be forwarded to the USPTO.
Flagging a post as an ACTION ITEM signals further research. Flagging SPAM and ABUSE helps to manage discussion.
Placing double brackets around a reference to a claim or prior art will create a hyperlink to the original
ex. [[claim 1]] and [[prior art 2]].
Facilitator's Comment 
Action Item Show without Noise
2
Susan Murray (6 months ago)
Teddi - I see you have some definite thoughts about this application. The Patent Examiner will need more detailed information. Do you know of some written publications which predate the filing date and explain the concepts covered here? Technical papers , product documentation, other patents or published patent applications would all be good sources. If you don't, who do you know that might have this information ... you can use "Invite a Reviewer" to pull others into the collaboration.1
Teddi Maranzano (6 months ago)
The problem this application is trying to address is how to move a workload (user and/or kernel) from a CPU without having to suspend the workload. The solution seems highly dependent upon the underlying hardware architecture. The application describes a computer system comprising several "boards" each comprising "one or more CPUs and some physical memory." Each domain is allocated "one or more system boards." With this architecture, when "it may become desirable to change the allocation of system boards to domains" or "that system board's physical pages needs to be relocated to the physical pages of another system board" the algorithm must be customized to the architecture. By contrast, the problem of moving a workload without suspending the workload has been solved by at least one vendor as early as 2003. That architecture overlays the physical memory/cpu layer with a virtual layer, and organizes the memory and CPUs into a pool. Unless the domain is configured for processor or memory affinity, the workload communicates with the virtual layer. This vendor's architecture provides dynamic deallocation of memory and/or CPU when predictive failure is detected. The workload is migrated to other resources within the pool and the failing component is marked down and removed from the pool.
