In some examples, a system generates a neural network comprising logical identifiers of compute resources. For executing the neural network, the system maps the logical identifiers to physical addresses of physical resources, and loads instructions of the neural network onto the physical resources, wherein the loading comprises converting the logical identifiers in the neural network to the physical addresses.

Embodiments of the present invention provide systems and methods for evaluating and weighting resource usage activity data. The system may establish a communicable link to a user device via a user application to receive resource activity data and historical data from one or more users or systems via multiple communication channels. The system may evaluate the historical data and determine evaluation criteria based on perceived chance of loss associated with particular metadata characteristics, and use the evaluation criteria as weighted metrics for determining an overall evaluation score for the user based on indication from resource activity data that the user has conducted resource transfers with entities or channels identified in the historical data.

Techniques are provided for generating host connectivity plans with load balancing and resiliency. One method comprises obtaining a number of storage system target ports needed for a given host; identifying available target ports in the storage system and an input-output (IO) target component associated with each available target port; and calculating the host connectivity plan until the host connectivity plan includes the obtained number of target ports by: (i) selecting at least one IO target component not already in the host connectivity plan that satisfies a resiliency policy and/or a load balancing policy; (ii) selecting at least one target port associated with the selected at least one IO target component and (iii) adding the selected at least one target port to the host connectivity plan. The resiliency policy may require connectivity without a single point of failure. The load balancing policy may specify that the IO target components serve a substantially equal IO load.

In some examples, a computing system receives an indication of an increased workload portion to be added to a workload of a storage system, the workload comprising buckets of operations of different characteristics. The computing system computes, based on quantities of operations of the different characteristics in the workload, factor values that indicate distribution of operations of the increased workload portion to the buckets of operations of the different characteristics, and distributes, according to the factor values, the operations of the increased workload portion into the buckets of operations of the different characteristics.

A method and system elects a master node from a plurality of nodes in a distributed system. A serverless elector function periodically outputs an election API call to a load balancer. The load balancer elects a master node from a plurality of candidate nodes each time the load balancer receives the election API call.

A transfer method is performed by an information processing apparatus. The method includes: selecting, based on a load status of the information processing apparatus, candidate transfer data that is among the received data and to be transferred to one or more other information processing apparatuses; selecting, based on load statuses of multiple other information processing apparatuses, one or more candidate transfer destination apparatuses among the multiple other information processing apparatuses as candidate transfer destinations of the data; determining, based on throughput between the information processing apparatus and the candidate transfer destination apparatuses, data to be transferred among the candidate transfer data, transfer destination apparatuses of the data to be transferred among the candidate transfer destination apparatuses, and the sizes of data groups including the data to be transferred; and transferring, to the transfer destination apparatuses determined for the determined data groups, the determined data to be transferred.

Example implementations relate to consensus protocols in a stretched network. According to an example, a distributed system includes continuously monitoring network performance and/or network latency among a cluster of a plurality of nodes in a distributed computer system. Leadership priority for each node is set based at least in part on the monitored network performance or network latency. Each node has a vote weight based at least in part on the leadership priority of the node. Each node's vote is biased by the node's vote weight. The node having a number of biased votes higher than a maximum possible number of votes biased by respective vote weights received by any other node in the cluster is selected as a leader node.

A method includes identifying, by a first instance of a service, a first number of data partitions of a data source to be processed by the service and a second number of instances of the service available to process the first number of data partitions. The method further includes separating the first number of data partitions into a first set of data partitions and a second set of data partitions in view of the second number of instances of the service, determining a target number of data partitions from the first set of data partitions to be claimed by each of the second number of instances of the service, and claiming, by the first instance of the service, the target number of data partitions from the first set of data partitions and up to one data partition from the second set of data partitions.

Various systems and methods are provided in which a replication process is initiated between a primary site and a recovery site, each having plurality of gateway appliances. Replication loads are evaluated for each given gateway appliance of the plurality of gateway appliances. If a determination is made that at least one gateway appliance of the plurality of gateway appliances is not overloaded, the plurality of gateway appliances are sorted based on replication loads respectively associated with each gateway appliance, and a determination is made as to whether a relative difference in replication loads between a gateway appliance having a highest replication load and a gateway appliance having a lowest replication load exceeds a difference threshold to determine whether the replication workloads between the gateway appliances should be rebalanced.

An example method for storing data includes providing a plurality of physical storage pools, each storage pool including a plurality of storage nodes coupled to a network. The method also includes mapping a partition of a plurality of partitions to a set of physical storage pools, where each physical storage pool of the set of physical storage pools is located in a different availability zone, and the storage nodes within an availability zone are subject to a correlated loss of access to stored data. The method further includes receiving a data management request over the network, the data management request being associated with a data object. The method also includes identifying a first partition of the plurality of partitions corresponding to the received data management request and manipulating the data object in the physical storage pools mapped to the first partition in accordance with the data management request.