摘要:
Provided are a system and a method for indexing high-dimensional data in parallel in a cluster environment. The system for indexing high-dimensional data in parallel in a cluster environment includes a Spill-tree creation means for creating a Spill-tree using an sampled N-dimensional feature vector, a feature vector division storage means for distributedly storing the N-dimensional feature vector in a terminal node of the Spill-tree, and a local signature creation means for creating and managing a local signature for the N-dimensional feature vector dispersed into each node of the Spill-tree.
摘要:
A system and method for providing streaming game software for a portable terminal are provided. The system includes a software producer, a service provider, a portable terminal, and a billing device. The software producer produces streaming software, and the service provider provides the software in a streaming mode. The portable terminal creates virtual computing resources by virtualizing and amplifying real computing resources for executing software and executes streaming software provided from the service provider based on the amplified virtual computing resource. The billing devices collects billing information, and calculates a cost according to a billing policy set based on the collected billing information.
摘要:
Provided is a method for providing personalized information lifecycle management service based on a personal data management policies. The method includes the steps of a) determining whether there is material change between a system for providing the information lifecycle management service and a first storage device connected to the system through a network; b) identifying the personal data management policies for the connected first storage device; c) identifying user's context information; d) performing data management based on the identified personal data management policies and context information; and e) updating the metadata with the execution result of data management.
摘要:
Provided is a stream data processing system and method for avoiding duplication of data process. The system including: an evaluation result storing unit for updating and storing a query condition evaluation result; a window evaluating unit for performing window evaluation; a data separating unit for separating data into new data and duplication input data; a reuse result extracting unit for receiving duplication input data from the data separating unit and extracting a query condition evaluation result; a query condition evaluating unit for receiving new data from the data separating unit, performing query condition evaluation and creating a query condition evaluation result; and a result organizing unit for receiving the query condition evaluation result, merging, outputting and transmitting the query condition evaluation result to the evaluation result storing unit.
摘要:
A packet detecting device used in a receiver employing PCI Express protocol is disclosed. The packet detecting device includes: a physical layer packet detecting unit for detecting a physical layer packet from a PCI express packet received and parallelized to 16 bit data through a physical deserializer; a descrambling unit for descrambling a physical layer packet from a PCI express packet received and parallelized to 16 bit data through a physical deserializer; a data link layer packet detecting unit for detecting a data link layer packet from a descrambled packet outputted from the descrambling unit; and a transaction layer packet detecting unit for detecting a transaction layer packet from a descrambled packet outputted from the descrambling unit.
摘要:
A cache consistency maintenance procedure to select one of a detection-based cache consistency maintenance procedure optimized for record-based locking and an avoidance-based consistency-based maintenance procedure optimized for table and block-based locking. To support the characteristic of DBMS in which table locking and record locking are consistent to access the same table, the two kinds of the consistency maintenance policies for the same block are processed by a single buffer load process and the two kinds of the consistency maintenance policies are consistent with each other to provide better configuration and performance.
摘要:
A concurrency control method for a high dimensional index structure that provides efficient concurrency control method for a high dimensional index structure, which performs reinsertion of certain objects to cope with node overflow. The concurrency controlled searching method includes the following steps. First, an entry is obtained from a queue storing the root node and an object relating to the entry is selected. Second, whether a logic sequence number (LSN) of a lower level node is larger than an expected LSN stored in the upper node is determined. Third, the process moves to a neighbor node of the lower level node if the LSN is bigger than an expected LSN stored in the upper node in the second step, selects a relating object, and performs from the second step repeatedly. Fourth, an object of a node of a level corresponding to the lower level node in a reinsertion table is selected when a search on an index tree is finished if the LSN is not bigger than the expected LSN stored in the upper node in the second step.