Dimension based load balancing

One or more computer-storage media storing computer-useable instructions that, when executed by a computing device, perform a method for load balancing scalable storage, comprising:

• receiving partitions of a plurality of partitions for assignment to servers of the scalable storage, the partitions having partition values of dimensions, each of the dimensions corresponding to one or more resources of the scalable storage;
• determining assignment plans for the assignment of the partitions to the servers of the scalable storage by bin packing the partition values of the dimensions into server bins of the servers, wherein the bin packing for different ones of the assignment plans utilize different assignment heuristics;
• selecting one of the assignment plans for execution on the scalable storage based on analyzing the partition values in accordance with assignments of the assignment plans;
• executing the selected one of the assignment plans on the scalable storage.

A computer-implemented method for load balancing a scalable storage, the method comprising:

• receiving partitions for assignment to servers of the scalable storage;
• generating assignment plans for the assignment of the partitions to the servers based on dimensional values of dimensions, each of the dimensions corresponding to one or more resources of the scalable storage, wherein the assignment plans are determined by different assignment heuristics;
• selecting one of the assignment plans for execution on the scalable storage based on analyzing the dimensional values in accordance with the assignments of the assignment plans;
• executing the selected one of the assignment plans on the scalable storage.

A computer-implemented method for load balancing a scalable storage, the method comprising:

• determining, for a triggered optimization module of a plurality of optimization modules, that at least one server of the scalable storage is over utilized on a dimension that corresponds to one or more resources of the scalable storage;
• optimizing the scalable storage based on an optimization routine, the optimization routine comprising:
• selecting a plurality of candidate operations that are potential load balancing operations associated with partitions assigned to the at least one server of the scalable storage;
• for a selected optimization module of the plurality of optimization modules that has a higher priority than the triggered optimization module, removing a candidate operation from the plurality of candidate operations that would diminish a modeled state of the scalable storage as quantified by an evaluation score of the selected optimization module;
• and
• determining at least one operation of the plurality of candidate operations that would improve the modeled state of the scalable storage with respect to a metric of the dimension on the at least one server as quantified by an evaluation score of the triggered optimization module, the metric quantifying utilization of the one or more resources;
• executing the at least one operation on the scalable storage.

One or more computer-storage media storing computer-useable instructions that, when executed by a computing device, perform a method for load balancing scalable storage, comprising:

• identifying a plurality of servers of the scalable storage that have a metric exceeding a cap value for at least one dimension of a plurality of dimensions, the metric being of a plurality of metrics that quantify utilization of one or more resources of the plurality of dimensions with respect to the plurality of servers;
• selecting at least one server of the plurality of servers based on a load metric of the at least server, the load metric being based on the plurality of metrics;
• providing a plurality of candidate operations, wherein the plurality of candidate operations are potential load balancing operations associated with partitions assigned to the at least one server;
• and
• selecting at least one operation of the plurality of candidate operations based on an evaluation score;
• executing the at least one operation on the scalable storage.

A computer-implemented method for load balancing a scalable storage, the method comprising:

• selecting at least one source server of a plurality of servers of the scalable storage that has a metric exceeding a cap value for at least one dimension of a plurality of dimensions, the metric being of a plurality of metrics that quantify utilization of one or more resources of the plurality of dimensions with respect to the plurality of servers;
• determining a plurality of destination servers for receiving partitions that are assigned to the at least one source server based on a load metric with respect to the plurality of servers, the load metric being based on the plurality of metrics;
• generating a plurality of candidate operations, wherein each the plurality of candidate operations are potential load balancing operations that correspond to at least one of the partitions assigned to the at least one source sever being received by at least one of the plurality of destination servers;
• selecting at least one operation of the plurality of candidate operations based on an evaluation score;
• executing the at least one operation on the scalable storage.