Patents by Inventor Mor Harchol-Balter
Mor Harchol-Balter has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 9052895Abstract: Systems, apparatuses, methods, and software that implement power budget allocation optimization algorithms in multi-processor systems, such as server farms. The algorithms are derived from a queuing theoretic model that minimizes the mean response time of the system to the jobs in the workload while accounting for a variety of factors. These factors include, but are not necessarily limited to, the type of power (frequency) scaling mechanism(s) available within the processors in the system, the power-to-frequency relationship(s) of the processors for the scaling mechanism(s) available, whether or not the system is an open or closed loop system, the arrival rate of jobs incoming into the system, the number of jobs within the system, and the type of workload being processed.Type: GrantFiled: April 7, 2011Date of Patent: June 9, 2015Assignees: International Business Machines, Carnegie Mellon UniversityInventors: Mor Harchol-Balter, Anshul Gandhi, Rajarshi Das, Jeffrey Kephart
-
Patent number: 8806018Abstract: A dynamic capacity management policy for multi-paralleled computing resources (e.g., application servers, virtual application servers, etc.) that includes one or more of a state-change component, a load-balancing component, and a robustness-control component. The state-change component delays the release (e.g., powering down of a physical server, removal from a virtual-server lease, etc.) of each computing resource for a set amount of time. The load-balancing component can work in conjunction with the state-change component to reduce the number of idle computing resources by distributing incoming requests in a manner that keeps the already-processing computing resources as full of requests as possible. The robustness-control component scales capacity as a function of the current number of requests within the system of computing resources to account for variations other than request rate, such as request size, reduced processor frequency, network slowdowns, etc., that affect processing capacity.Type: GrantFiled: March 30, 2012Date of Patent: August 12, 2014Assignees: Carnegie Mellon University, Intel CorporationInventors: Mor Harchol-Balter, Anshul Gandhi, Varun Gupta, Michael Kozuch
-
Patent number: 8589709Abstract: Processor-management techniques that purposely alternate a processor between an operating state and a non-operating state while the processor is executing the workload. The techniques leverage the “ultra-low-power” non-operating states of many processors to provide predictable power and/or frequency control of the processor. These techniques can provide better performance than known clock-throttling and dynamic voltage and frequency scaling schemes for controlling processors.Type: GrantFiled: July 22, 2010Date of Patent: November 19, 2013Assignee: Carnegie Mellon UniversityInventors: Mor Harchol-Balter, Anshul Gandhi
-
Patent number: 8537717Abstract: In distributed networks of cooperating nodes, it is useful to perform resource discovery in a manner that is efficient but that also minimizes communication complexity. A system and method in which nodes in a network efficiently are provided with information about the presence of, and other information about, other nodes in the network provides tangible benefits. In general, in one aspect, a system and method according to the invention features a distributed method for communicating information among a plurality of nodes. The method includes communicating from the first node to the second node information about the first node and nodes that the first node is aware of. The method further includes adding or merging, by the second node, the information about the first node and nodes that the first node is aware of with information about nodes that the second node is aware of.Type: GrantFiled: August 26, 2010Date of Patent: September 17, 2013Assignee: Massachusetts Institute of TechnologyInventors: Mor Harchol-Balter, F. Thomson Leighton, Daniel Lewin
-
Publication number: 20120254444Abstract: A dynamic capacity management policy for multi-paralleled computing resources (e.g., application servers, virtual application servers, etc.) that includes one or more of a state-change component, a load-balancing component, and a robustness-control component. The state-change component delays the release (e.g., powering down of a physical server, removal from a virtual-server lease, etc.) of each computing resource for a set amount of time. The load-balancing component can work in conjunction with the state-change component to reduce the number of idle computing resources by distributing incoming requests in a manner that keeps the already-processing computing resources as full of requests as possible. The robustness-control component scales capacity as a function of the current number of requests within the system of computing resources to account for variations other than request rate, such as request size, reduced processor frequency, network slowdowns, etc., that affect processing capacity.Type: ApplicationFiled: March 30, 2012Publication date: October 4, 2012Applicant: CARNEGIE MELLON UNIVERSITYInventors: Mor Harchol-Balter, Anshul Gandhi, Varun Gupta, Michael Kozuch
-
Publication number: 20120084580Abstract: Systems, apparatuses, methods, and software that implement power budget allocation optimization algorithms in multi-processor systems, such as server farms. The algorithms are derived from a queuing theoretic model that minimizes the mean response time of the system to the jobs in the workload while accounting for a variety of factors. These factors include, but are not necessarily limited to, the type of power (frequency) scaling mechanism(s) available within the processors in the system, the power-to-frequency relationship(s) of the processors for the scaling mechanism(s) available, whether or not the system is an open or closed loop system, the arrival rate of jobs incoming into the system, the number of jobs within the system, and the type of workload being processed.Type: ApplicationFiled: April 7, 2011Publication date: April 5, 2012Applicants: International Business Machines, Carnegie Mellon UniversityInventors: Mor Harchol-Balter, Anshul Gandhi, Rajarshi Das, Jeffrey Kephart
-
Publication number: 20110106887Abstract: In distributed networks of cooperating nodes, it is useful to perform resource discovery in a manner that is efficient but that also minimizes communication complexity. A system and method in which nodes in a network efficiently are provided with information about the presence of, and other information about, other nodes in the network provides tangible benefits. In general, in one aspect, a system and method according to the invention features a distributed method for communicating information among a plurality of nodes. The method includes choosing, by a first node, one second node from information about nodes that the first node is aware of. The method further includes communicating from the first node to the second node information about the first node and nodes that the first node is aware of. The method further includes adding or merging, by the second node, the information about the first node and nodes that the first node is aware of with information about nodes that the second node is aware of.Type: ApplicationFiled: August 26, 2010Publication date: May 5, 2011Inventors: Mor Harchol-Balter, F. Thomson Leighton, Daniel Lewin
-
Publication number: 20110022868Abstract: Processor-management techniques that purposely alternate a processor between an operating state and a non-operating state while the processor is executing the workload. The techniques leverage the “ultra-low-power” non-operating states of many processors to provide predictable power and/or frequency control of the processor. These techniques can provide better performance than known clock-throttling and dynamic voltage and frequency scaling schemes for controlling processors.Type: ApplicationFiled: July 22, 2010Publication date: January 27, 2011Applicant: CARNEGIE MELLON UNIVERSITYInventors: Mor Harchol-Balter, Anshul Gandhi
-
Patent number: 7808926Abstract: In distributed networks of cooperating nodes, it is useful to perform resource discovery in a manner that is efficient but that also minimizes communication complexity. A system and method in which nodes in a network efficiently are provided with information about the presence of, and other information about, other nodes in the network provides tangible benefits. In general, in one aspect, a system and method according to the invention features a distributed method for communicating information among a plurality of nodes. The method includes choosing, by a first node, one second node from information about nodes that the first node is aware of. The method further includes communicating from the first node to the second node information about the first node and nodes that the first node is aware of. The method further includes adding or merging, by the second node, the information about the first node and nodes that the first node is aware of with information about nodes that the second node is aware of.Type: GrantFiled: October 2, 2007Date of Patent: October 5, 2010Assignee: Massachusetts Institute of TechnologyInventors: Mor Harchol-Balter, F. Thomson Leighton, Daniel Lewin
-
Publication number: 20080212492Abstract: In distributed networks of cooperating nodes, it is useful to perform resource discovery in a manner that is efficient but that also minimizes communication complexity. A system and method in which nodes in a network efficiently are provided with information about the presence of, and other information about, other nodes in the network provides tangible benefits. In general, in one aspect, a system and method according to the invention features a distributed method for communicating information among a plurality of nodes. The method includes choosing, by a first node, one second node from information about nodes that the first node is aware of. The method further includes communicating from the first node to the second node information about the first node and nodes that the first node is aware of. The method further includes adding or merging, by the second node, the information about the first node and nodes that the first node is aware of with information about nodes that the second node is aware of.Type: ApplicationFiled: October 2, 2007Publication date: September 4, 2008Applicant: Massachusetts Institute of TechnologyInventors: Mor Harchol-Balter, F. Thomson Leighton, Daniel Lewin
-
Patent number: 7295523Abstract: In distributed networks of cooperating nodes, it is useful to perform resource discovery in a manner that is efficient but that also minimizes communication complexity. A system and method in which nodes in a network efficiently are provided with information about the presence of, and other information about, other nodes in the network provides tangible benefits. In general, in one aspect, a system and method according to the invention features a distributed method for communicating information among a plurality of nodes. The method includes communicating from the first node to the second node information about the first node and nodes that the first node is aware of. The method further includes adding or merging, by the second node, the information about the first node and nodes that the first node is aware of with information about nodes that the second node is aware of.Type: GrantFiled: May 23, 2000Date of Patent: November 13, 2007Assignee: Massachusetts Institute of TechnologyInventors: Mor Harchol-Balter, F. Thomson Leighton, Daniel Lewin
-
Patent number: 6223205Abstract: A distributed server system is disclosed which includes a load balancer and a plurality of host processors. The load balancer receives requests for service and distributes task assignments among the plurality of processors based upon the amount of work associated with the respective requests for service. More specifically, each host processor services requests for service within a predefined task size interval and the load balancer assigns to each host processor only those requests for service which involve task sizes within the particular task size interval associated with the respective processor. In the foregoing manner, the variability of the task sizes assigned to any given host processor is minimized and performance of the distributed server system is improved.Type: GrantFiled: February 13, 1998Date of Patent: April 24, 2001Inventors: Mor Harchol-Balter, Mark E. Crovella