Abstract: According to an aspect, a method for key-value storage includes determining whether a data size of a first set comprising data and a key stored in a first area or a second area of a first storage unit is larger than a threshold determined for the first area or the second area in which the first set is stored, as a first execution condition. The method further includes determining whether a degree of collisions of a hash value indicating a storage location of the first set in the first area of the first storage unit is greater than a predetermined threshold, as a second execution condition. When the set satisfies either of the first or second execution conditions, the method includes the first set in the second area.

Abstract: According to an aspect, a method for key-value storage includes determining whether a data size of a first set comprising data and a key stored in a first area or a second area of a first storage unit is larger than a threshold determined for the first area or the second area in which the first set is stored, as a first execution condition. The method further includes determining whether a degree of collisions of a hash value indicating a storage location of the first set in the first area of the first storage unit is greater than a predetermined threshold, as a second execution condition. When the set satisfies either of the first or second execution conditions, the method includes the first set in the second area.

Abstract: According to an aspect, a key-value store (KVS) system includes a data management unit that stores a data KVS storing a pair of a data KVS key including information on a storage location of application data to be an access target object and the application data; and a key KVS storing a pair of an application key and the data KVS key. The data KVS includes a normal partition in which a size of a record for storing one pair is a predetermined specific size; and a special partition in which the size of the record for storing one pair is a size set according to a data size of the pair to be stored. A data relocation unit relocates a pair of a relocation target object to the special partition having the record size suitable for the data size of the pair.

Abstract: According to an aspect, a key-value store (KVS) system includes a data management unit that stores a data KVS storing a pair of a data KVS key including information on a storage location of application data to be an access target object and the application data; and a key KVS storing a pair of an application key and the data KVS key. The data KVS includes a normal partition in which a size of a record for storing one pair is a predetermined specific size; and a special partition in which the size of the record for storing one pair is a size set according to a data size of the pair to be stored. A data relocation unit relocates a pair of a relocation target object to the special partition having the record size suitable for the data size of the pair.

Abstract: A task count controller, a task count control method, and a computer program capable of dynamically controlling the number of tasks that can be processed in parallel simultaneously without increasing computational load are provided. When a plurality of tasks are to be executed simultaneously in parallel processing, the number of tasks that can be executed simultaneously is controlled. The tasks to be executed simultaneously are added in units of a predetermined number of tasks and the throughput in one unit of work is measured for each task every time the tasks are added. The total sum of the measured throughputs is calculated, and it is determined whether the calculated total sum of throughputs is more than the total sum of throughputs immediately before the predetermined number of tasks are added.

Abstract: A task count controller, a task count control method, and a computer program capable of dynamically controlling the number of tasks that can be processed in parallel simultaneously without increasing computational load are provided. When a plurality of tasks are to be executed simultaneously in parallel processing, the number of tasks that can be executed simultaneously is controlled. The tasks to be executed simultaneously are added in units of a predetermined number of tasks and the throughput in one unit of work is measured for each task every time the tasks are added. The total sum of the measured throughputs is calculated, and it is determined whether the calculated total sum of throughputs is more than the total sum of throughputs immediately before the predetermined number of tasks are added.