公开(公告)号：US07933938B2

公开(公告)日：2011-04-26

申请号：US11860652

申请日：2007-09-25

申请人： Myung Hoon Cha ,  Hong Yeon Kim ,  June Kim ,  Myung Joon Kim

发明人： Myung Hoon Cha ,  Hong Yeon Kim ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F12/00

CPC分类号： G06F17/301 ,  G06F17/30147

摘要： A file storage system, a file storing method, and a file searching method. In the file storing method, a single file to be stored is classified into at least one or more objects, and the classified objects are classified into pages with a serial number. One storage device to record the file is selected among a plurality of storage devices connected through a network. One object is allocated to the selected storage device. Pages configuring the file of the object allocated to the selected storage device are stored. When all pages of the file are not stored, the remaining pages are stored by allocating new object to a new storage device. Metadata for managing information about the storage of the file is set. Accordingly, the file can be stored over several disks. In addition, the overall system performance can be increased by setting the file information in the object metadata simply using the least resources.

摘要翻译： 文件存储系统，文件存储方法和文件搜索方法。 在文件存储方法中，要存储的单个文件被分类为至少一个或多个对象，分类对象被分类为具有序列号的页面。 在通过网络连接的多个存储装置中选择用于记录文件的一个存储装置。 一个对象被分配给所选择的存储设备。 存储配置分配给所选存储设备的对象的文件的页面。 当文件的所有页面都未被存储时，通过将新对象分配给新的存储设备来存储剩余的页面。 设置用于管理有关文件存储信息的元数据。 因此，文件可以存储在几个磁盘上。 此外，通过仅使用最少的资源将文件信息设置在对象元数据中，可以提高整个系统性能。

公开(公告)号：US07676607B2

公开(公告)日：2010-03-09

申请号：US11636046

申请日：2006-12-08

申请人： Byung Kwon Jung ,  Young Kyun Kim ,  June Kim ,  Myung Joon Kim

发明人： Byung Kwon Jung ,  Young Kyun Kim ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F3/00 ,  G06F13/00

CPC分类号： H04L69/16 ,  G06F3/0658 ,  G06F3/0689 ,  H04L69/12 ,  H04L69/161

摘要： Provided is a hardware acceleration apparatus for an iSCSI target system including a local memory storing data transmitted between a disk storage device and a network system, a TOE controller performing TCP/IP stack offloading on a TCP/IP packet of data stored in the local memory, a network controller allowing data offloaded from the local memory to be transmitted to the network system using an iSCSI protocol or allowing data received from the network system to be stored in the local memory, and a disk controller storing and withdrawing data between the disk storage device and the local memory.

摘要翻译： 提供了一种用于iSCSI目标系统的硬件加速装置，其包括存储在盘存储装置和网络系统之间传输的数据的本地存储器，在存储在本地存储器中的数据的TCP / IP分组上执行TCP / IP堆栈卸载的TOE控制器 允许从本地存储器卸载数据的网络控制器使用iSCSI协议向网络系统发送或允许从网络系统接收到的数据存储在本地存储器中;以及磁盘控制器，其在磁盘存储器之间存储和取出数据 设备和本地内存。

公开(公告)号：US20070143340A1

公开(公告)日：2007-06-21

申请号：US11634275

申请日：2006-12-05

申请人： Sang Lee ,  Young Kim ,  June Kim ,  Myung Joon Kim

发明人： Sang Lee ,  Young Kim ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F17/00

CPC分类号： G06F12/0875 ,  G06F12/0804 ,  Y10S707/955

摘要： A cache coherency maintenance system and method of a user file manager in an object-based storage system are provided. According to the cache coherence maintenance system and method, by maintaining the coherency of a writeback cache based on time using a timer, metadata and data information are efficiently managed. Accordingly, even when accesses by users increase, the cost of maintaining the coherency of the cache is distributed to respective user file managers. In this way the performance of the entire object-based storage system is enhanced and the consistency and efficiency of information increase.

摘要翻译： 提供了一种基于对象的存储系统中的用户文件管理器的高速缓存一致性维护系统和方法。 根据缓存一致性维护系统和方法，通过使用定时器基于时间维持回写高速缓存的一致性，有效地管理元数据和数据信息。 因此，即使当用户的访问增加时，将高速缓存的一致性维持的成本分配给相应的用户文件管理器。 以这种方式，整个基于对象的存储系统的性能得到提高，信息的一致性和效率也提高了。

公开(公告)号：US20050138048A1

公开(公告)日：2005-06-23

申请号：US10821870

申请日：2004-04-12

申请人： Ki Sung Jin ,  June Kim ,  Mi Young Lee ,  Myung Joon Kim

发明人： Ki Sung Jin ,  June Kim ,  Mi Young Lee ,  Myung Joon Kim

IPC分类号： G06F12/16 ,  G06F7/00 ,  G06F17/30

CPC分类号： G06F16/83

摘要： Disclosed is an XML database duplicating apparatus and method capable of providing an effective application service without interruption at a low cost in the diverse Internet environments by copying an XML document of a database system mounting an XML processing technique to another remote server. When the XML document stored in a database of the original server is copied to databases of copied servers, the structure and attribute information included in the XML document are maintained as they are, and thus the respective copied servers can support the searches based on the contents and the attribute provided by an XML query language and disperse users' queries concentrated to the original server to the respective copied servers, so that the service request for application programs can be processed.

摘要翻译： 公开了一种XML数据库复制装置和方法，其能够通过将安装XML处理技术的数据库系统的XML文档复制到另一个远程服务器，以不间断的方式在不同的互联网环境中以低成本提供有效的应用服务。 当存储在原始服务器的数据库中的XML文档被复制到复制的服务器的数据库时，包括在XML文档中的结构和属性信息原样维护，并且因此各个复制的服务器可以基于内容支持搜索 以及由XML查询语言提供的属性，并将用户的查询集中在原始服务器上，并分发给相应的复制服务器，以便可以处理应用程序的服务请求。

公开(公告)号：US06631385B2

公开(公告)日：2003-10-07

申请号：US09497136

申请日：2000-02-03

申请人： Jang Sun Lee ,  June Kim ,  Hun Soon Lee ,  Myung Joon Kim ,  Gil Rok Oh ,  Jae Soo Yoo ,  Seok Il Song

发明人： Jang Sun Lee ,  June Kim ,  Hun Soon Lee ,  Myung Joon Kim ,  Gil Rok Oh ,  Jae Soo Yoo ,  Seok Il Song

IPC分类号： G06F1730

CPC分类号： G06F11/1474 ,  Y10S707/99933 ,  Y10S707/99953

摘要： A recovery method for a high-dimensional index structure is disclosed, in which a reinsert operation is employed based on ARIES (algorithm for recovery and isolation exploiting semantics) and a page-oriented redo and a logical undo. Further, a recording medium on which a program for carrying out the above method is recorded is disclosed, the program being readable by a computer. The recovery method for a high-dimensional index structure employing a reinsert operation according to the present invention includes the following steps. At a first step, an entry is inserted into a node, a minimum bounding region is adjusted, an overflow is processed, and a log record is stored. At a second step, the log record thus stored is recovered.

摘要翻译： 公开了一种高维度索引结构的恢复方法，其中基于ARIES（用于恢复和隔离开发语义的算法）和面向页面的重做和逻辑撤销来采用重新插入操作。 此外，公开了记录有用于执行上述方法的程序的记录介质，该程序可由计算机读取。 根据本发明的采用重新插入操作的高维索引结构的恢复方法包括以下步骤。 在第一步，将条目插入节点，调整最小边界区域，处理溢出，并存储日志记录。 在第二步，恢复如此存储的日志记录。

公开(公告)号：US07849355B2

公开(公告)日：2010-12-07

申请号：US11948918

申请日：2007-11-30

申请人： Sang Min Lee ,  June Kim ,  Myung Joon Kim

发明人： Sang Min Lee ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F11/00

CPC分类号： G06F11/2082 ,  G06F11/2087

摘要： A distributed object sharing system and a method thereof are provided. In a file server writing method for storing an object in a RAID (redundant array of inexpensive disks) apparatus having a plurality of storage apparatuses, the file server writing method includes: a first process of, when an object writing command for a specific object is generated during error recovery in an erroneous storage apparatus, checking a recovery state of the object with reference to recovery state information provided from a metadata server; a second process of, when the recovery of the object is completed, performing the object writing command on a normal storage apparatus and a recovery storage apparatus; a third process of, when the recovery of the object is proceeding, receiving the recovery state information from the metadata server again, checking the recovery state of the object again, and determining whether or not to perform the second process; a fourth process of, when the second process is determined not to performed, performing the object writing command on only the normal storage apparatus and receiving the recovery state information from the metadata server again; and a fifth process of checking the recovery state of the object with reference to the received recovery state information and, when the recovery of the object is completed, additionally performing the object writing command on the recovery storage apparatus. Accordingly, it is possible to prevent data loss and perform I/O command during error recovery.

摘要翻译： 提供了一种分布式对象共享系统及其方法。 在用于将对象存储在具有多个存储装置的RAID（廉价磁盘的冗余阵列）装置的文件服务器写入方法中，文件服务器写入方法包括：第一处理，当对特定对象的对象写入命令为 参照从元数据服务器提供的恢复状态信息，检查对象的恢复状态; 当对象的恢复完成时，在正常存储装置和恢复存储装置上执行对象写入命令的第二处理; 第三过程，当对象的恢复正在进行时，再次从元数据服务器接收恢复状态信息，再次检查对象的恢复状态，并确定是否执行第二处理; 第四处理当第二处理被确定为不执行时，仅在普通存储装置上执行对象写入命令，并再次从元数据服务器接收恢复状态信息; 以及参照所接收的恢复状态信息来检查对象的恢复状态的第五处理，并且当对象的恢复完成时，在恢复存储装置上附加执行对象写入命令。 因此，可以防止数据丢失，并在错误恢复期间执行I / O命令。

公开(公告)号：US07797275B2

公开(公告)日：2010-09-14

申请号：US11634275

申请日：2006-12-05

申请人： Sang Min Lee ,  Young Kyun Kim ,  June Kim ,  Myung Joon Kim

发明人： Sang Min Lee ,  Young Kyun Kim ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F7/00

CPC分类号： G06F12/0875 ,  G06F12/0804 ,  Y10S707/955

摘要： A cache coherency maintenance system and method of a user file manager in an object-based storage system are provided. According to the cache coherence maintenance system and method, by maintaining the coherency of a writeback cache based on time using a timer, metadata and data information are efficiently managed. Accordingly, even when accesses by users increase, the cost of maintaining the coherency of the cache is distributed to respective user file managers. In this way the performance of the entire object-based storage system is enhanced and the consistency and efficiency of information increase.

摘要翻译： 提供了一种基于对象的存储系统中的用户文件管理器的高速缓存一致性维护系统和方法。 根据缓存一致性维护系统和方法，通过使用定时器基于时间维持回写高速缓存的一致性，有效地管理元数据和数据信息。 因此，即使当用户的访问增加时，将高速缓存的一致性维持的成本分配给相应的用户文件管理器。 以这种方式，整个基于对象的存储系统的性能得到提高，信息的一致性和效率也提高了。

公开(公告)号：US07596659B2

公开(公告)日：2009-09-29

申请号：US11634674

申请日：2006-12-06

申请人： Sang Min Lee ,  Young Kyun Kim ,  June Kim ,  Myung Joon Kim

发明人： Sang Min Lee ,  Young Kyun Kim ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F12/06

CPC分类号： G06F3/0659 ,  G06F3/061 ,  G06F3/0667 ,  G06F3/0689

摘要： A method of balanced object striping and a system therefor are provided. The object striping method includes: transmitting a data write request and receiving an object identifier if the data write request is a first write request regarding the data; determining a first object storage target to store the data by using the object identifier; and striping the data in a plurality of object storage targets starting from the first object storage target to store the data.

摘要翻译： 提供了一种平衡对象条纹化的方法及其系统。 对象条带化方法包括：如果数据写请求是关于数据的第一写请求，则发送数据写请求并接收对象标识符; 确定通过使用所述对象标识符来存储所述数据的第一对象存储目标; 以及从第一对象存储目标开始的多个对象存储目标中的数据分割以存储数据。

公开(公告)号：US20080133967A1

公开(公告)日：2008-06-05

申请号：US11948918

申请日：2007-11-30

申请人： Sang Min Lee ,  June Kim ,  Myung Joon Kim

发明人： Sang Min Lee ,  June Kim ,  Myung Joon Kim

IPC分类号： G06F11/14

CPC分类号： G06F11/2082 ,  G06F11/2087

摘要： A distributed object sharing system and a method thereof are provided. In a file server writing method for storing an object in a RAID (redundant array of inexpensive disks) apparatus having a plurality of storage apparatuses, the file server writing method includes: a first process of, when an object writing command for a specific object is generated during error recovery in an erroneous storage apparatus, checking a recovery state of the object with reference to recovery state information provided from a metadata server; a second process of, when the recovery of the object is completed, performing the object writing command on a normal storage apparatus and a recovery storage apparatus; a third process of, when the recovery of the object is proceeding, receiving the recovery date information from the metadata server again, checking the recovery state of the object again, and determining whether or not to perform the second process; a fourth process of, when the second process is determined not to performed, performing the object writing command on only the normal storage apparatus and receiving the recovery state information from the metadata server again; and a fifth process of checking the recovery state of the object with reference to the received recovery date information and, when the recovery of the object is completed, additionally performing the object writing command on the recovery storage apparatus. Accordingly, it is possible to prevent data loss and perform I/O command during error recovery.

摘要翻译： 提供了一种分布式对象共享系统及其方法。 在用于将对象存储在具有多个存储装置的RAID（廉价磁盘的冗余阵列）装置的文件服务器写入方法中，文件服务器写入方法包括：第一处理，当对特定对象的对象写入命令为 参照从元数据服务器提供的恢复状态信息，检查对象的恢复状态; 当对象的恢复完成时，在正常存储装置和恢复存储装置上执行对象写入命令的第二处理; 第三过程，当对象的恢复正在进行时，再次从元数据服务器接收恢复日期信息，再次检查对象的恢复状态，并确定是否执行第二处理; 第四处理当第二处理被确定为不执行时，仅在普通存储装置上执行对象写入命令，并再次从元数据服务器接收恢复状态信息; 以及参照所接收的恢复日期信息来检查对象的恢复状态的第五处理，并且当对象的恢复完成时，在恢复存储装置上另外执行对象写入命令。 因此，可以防止数据丢失，并在错误恢复期间执行I / O命令。

公开(公告)号：US20070180299A1

公开(公告)日：2007-08-02

申请号：US11605181

申请日：2006-11-28

申请人： Kap Dong Kim ,  Song Woo Sok ,  Hag Young Kim ,  Myung Joon Kim

发明人： Kap Dong Kim ,  Song Woo Sok ,  Hag Young Kim ,  Myung Joon Kim

IPC分类号： G06F11/00

CPC分类号： G06F11/1076

摘要： A method of data placement and control in a block-divided distributed disk array is provided. At first, data to store is divided into logical blocks, and each of the divided logical blocks is further divided into a plurality of data blocks. Then, a parity block is created through performing an XOR operation on the data blocks of each logical block in a unit of a strip, and the parity block is arranged at the first block of a strip unit. Then, parity blocks are arranged at a next block of a previous logical block's the last data block through a left/right shifting for uniformly arranging the parity blocks through out all disks in order to fill an empty block between the logical blocks.

摘要翻译： 提供了一种块分散分布式磁盘阵列中的数据放置和控制方法。 首先，要存储的数据被划分为逻辑块，并且每个划分的逻辑块被进一步划分为多个数据块。 然后，通过以条为单位对每个逻辑块的数据块执行异或运算来创建奇偶校验块，并且奇偶校验块被布置在条带单元的第一块处。 然后，通过左/右移位将奇偶校验块布置在先前逻辑块的最后数据块的下一个块上，以将奇偶校验块均匀排列出所有磁盘，以便填充逻辑块之间的空块。