Abstract: The present invention provides a recovery method using extendible hashing-based cluster logs in a shared-nothing spatial database cluster, which eliminates the duplication of cluster logs required for cluster recovery in a shared-nothing database cluster, so that recovery time is decreased, thus allowing the shared-nothing spatial database cluster system to continuously provide stable service. In the recovery method, if a failure occurs in a predetermined node, a second node in a group, including the node, records cluster logs in main memory on the basis of extendible hashing. If the node that has failed recovers itself using a local log, the second node in the group transmits cluster logs in packets to a recovery node that is the failed node. If the recovery node reflects the received cluster logs and maintains consistency with other nodes in the group, the recovery node resumes normal service.

Abstract: A liquid crystal display device for removing residual charge includes a plurality of data lines and a plurality of gate lines arranged along vertical and horizontal directions, respectively, on a transparent substrate, a source driver for supplying data voltage to the data lines, a gate driver for supplying gate voltage to the gate lines, and a plurality of common voltage lines connected to the data lines and the gate lines through a plurality of static electricity preventing units, wherein at least one of the static electricity preventing units is directly connected to the source driver.

Abstract: In order to prevent exposure mismatch on a boundary between exposure regions that causes pattern connection defects (including stitch defects), exposure is performed twice or more on a whole exposure region of a glass substrate. The exposure method includes aligning a reticle in a scanning direction, exposing the reticle pattern onto the glass substrate, moving the glass substrate one-half of the width of the reticle, and exposing an exposure area twice by repeating the exposing and moving steps.

Abstract: There is disclosed a liquid crystal display that is adaptive for improving picture quality by eliminating residual DC components within a horizontal electric field type liquid crystal display panel. A liquid crystal display device according to an embodiment of the present invention includes a substrate defined as a display part and a non-display part; a gate line formed on the substrate; a common line substantially parallel to the gate line; a data line crossing the gate line and the common line while being insulated, to define a pixel area; and at least one capacitor located in the non-display part and connected to at least one of the gate line, the common line and the data line for storing a remaining component of the display part and eliminating the stored component.

Abstract: The present invention relates to a recovery function from a system failure such as a power failure and a media failure such as a breakage of disk at a transaction processing system of a plurality of user environment, and is applied in carrying out a checkpoint to reduce an amount of work required for failure recovery, and particularly to a fuzzy checkpoint method which is a method of enhancing a system performance by not interrupting other transactions ever for a moment at the time of checkpoint. According to the present invention, the occurrance of dangling transaction is prevented by deleting the corresponding transaction entry from the transaction table during performing of the redo step, for the transactions terminated between the point of time when the checkpoint start log record of the last completed checkpoint is logged and the point of time when the checkpoint end log record is logged.

Abstract: The present invention relates to a latch structure for effectively controlling the users' simultaneous accesses to a critical section in a system supporting a multi-user environment and a method for managing the same. The latch is a data structure, having a mutual exclusion variable and a pointer to a request queue, used in providing a schedule for user requests. It is an object of the present invention to provide a latch structure and a method for managing the same which are simple and are capable of improving the concurrence of the users' accesses to a critical section through the implementation of a complete latch satisfying all of the conditions required by the critical section. The concurrence for the multi-user system and the performance of the entire system can be greatly improved.

Abstract: A method for carrying out a check point for preventing a dangling transaction occurrence. The method uses a transaction table initialization technique performed during a recovery function. The method is performed by avoiding interruption of transactions occurring at the time the check point is carried out and is performed by reducing the amount of work required at the time of performing a recovery function from a failure detected in the transaction processing system. The occurrence of a dangling transaction is prevented by using a transaction table initialization technique at an analysis step for the transaction terminated between a time interval commencing when a check point start log record of a prior completed check point is logged and ending at a point of time when the check point end log record is logged.

Abstract: A buffer management method using a buffer locking mechanism in a storage system for supporting a multi-user environment. The buffer hash anchor is acquired with no exclusive latch while allocation of the empty buffer is required to read a desired page into the buffer using a buffer locking chain including buffer locking entries in the form of a list. Page locking is executed only when a buffer associated with a desired page does not exist in the buffer frame chain. Accordingly, there is no interference with transactions requiring access to other pages. It is also possible to prevent the buffer hash anchor from using an exclusive latch for a lengthy period of time due to other transactions concurrently requiring allocation of empty buffers. Therefore, there is no degradation in performance of the entire system. Since the page locking is executed only for a page read from the disk, the overhead required for the page locking is minimized.