Electrical computers and digital processing systems: multicomput – Multicomputer data transferring via shared memory
Reexamination Certificate
1999-09-20
2003-09-30
Sheikh, Ayaz (Department: 2155)
Electrical computers and digital processing systems: multicomput
Multicomputer data transferring via shared memory
C709S201000, C709S203000, C709S214000, C709S215000, C709S216000, C709S217000, C709S218000, C709S219000, C707S793000, C707S793000, C707S793000
Reexamination Certificate
active
06629132
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to indexing information stored in a computer system, and more particularly to indexing data stored in a cache of a directory proxy server.
BACKGROUND OF THE INVENTION
It is common for computer users (“clients”) interconnected by an institutional intranet or local area network to gain access to various remote (database) server sites via an internetwork of computers, such as the well-known Internet communications network. It is also common in network applications to provide a so-called proxy server that links to the internetwork. A proxy server accesses frequently requested data from the remote servers and stores it locally to effectively speed-up access and reduce the download time of future requests for the data. In response to a request from an application executing on a client, the proxy server attempts to fulfill that request from its local storage; if it cannot, the proxy server forwards the request over the internetwork to a server that can satisfy the request. The server then responds by transferring a stream of data to the proxy server, which stores and forwards the data onto the client.
The term “client” is also used to refer to a computer used by a person, the “user”. Accordingly, the user's computer is referred to as the “client computer”.
The requests issued from the client and proxy server to the server conform to a conventional protocol, such as the lightweight directory access protocol (LDAP). Specifically, the LDAP protocol provides a client-server communication arrangement to access a directory service over a Transmission Control Protocol/Internet Protocol (TCP/IP) network. Examples of a directory service include the NetWare Directory Services (NDS) from Novell, Inc. and the X.500 directory service. Novell's Directory Access Protocol (NDAP) is a gateway on NDS that conforms with LDAP. NDS, X.500 and the LDAP protocol are well-known and described in the following documents: Novell Directory Services Internals Overview; Technical Overview of Directory Services Using the X.500 Protocol, RFC 1309; X.500 Lightweight Directory Access Protocol, RFC 1487; Lightweight Directory Access Protocol (v3), RFC 2251.
A directory differs from a database in an essential characteristic, a directory is designed for ease of changing the data stored therein on a dynamic basis. In ordinary data base design, the data is stored in fields of tables, and is accessed and written to and read from using a designated protocol. To change data in a database requires both deleting the data presently there and writing in desired new data. Both the deleting and writing are accomplished by using the command structure of the protocol.
In contrast, a directory is architected so that the access protocol permits easy access to changing data stored in the directory. Protocols for dynamically changing data stored in a directory are designed to make dynamical changes to the data easy and able to be accomplished with a minimum of steps executed by the user, or his/her client computer. An example of directory operation and protocol is given in the Lightweight Directory Access Protocol (LDAP).
The LDAP protocol is described in many books, in particular in the following two books: the first book, by Timothy A. Howes, Mark C. Smith, and Gordon S. Good entitled Understanding and Deploying LDAP Directory Services, published by Macmillan Technical Publishing, Copyright date 1999; and second book, by Timothy A. Howes and Mark C. Smith entitled LDAP, Programming Directory Enabled Applications with Lightweight Directory Access Protocol, published by Macmillan Technical Publishing, Copyright date 1997, and all disclosures of both books are incorporated herein by reference.
A difference between a directory and a database can be expressed by the statement that a directory can include a database, but a database ordinarily cannot include a directory. A reason is that data may be stored in a directory much as it is stored in a database, but the access to a directory for dynamic changes in the stored data is better than access to a database. In the following discussion attention will be primarily directed to directories. However, as is clear from this discussion, a database could also be used in the discussion, with the exception that to use a database would make access for dynamic changes in the data more cumbersome.
In this document, the conventional protocol used to issue requests from a client is a lightweight directory access protocol (LDAP) and the source server used to store data is an LDAP or NDAP/NDS server. The predicate proxy server stores (“caches”) data retrieved from the server and further builds dynamic indexes for searching the cached data stored on the proxy cache. Notably, searching and storage of data on the proxy server is based on the predicate generated by the predicate logic core of the proxy server.
Any database management system may be used in the following description and used in the practice of the following invention. However, because of ease of reference, any database system, and any directory service, will be referred to as an “LDAP” directory service, whether or not it uses the LDAP protocol. That is, the present discussion is not limited to any specific protocol utilized by standard LDAP Lightweight Directory Access Protocol, even though the terminology “LDAP server” is used to refer to any electronically stored database.
The variants (types) of data stored in the LDAP (and NDAP and any directories using any other protocol) directories are typically small to make it easier for applications to directly access the data with a fully-qualified distinguished name; a distinguished name is a technique (similar to the Domain Naming System) for accessing data uniquely within a directory store. However, as the amount of data types stored in an LDAP/NDAP directory increases, it becomes increasingly difficult for an application and associated programs to access all the data and know about all their respective types. The directory may, for instance, contain different types (categories) of data such as printer identifiers (IDs), electronic mail (e-mail) addresses and Internet Protocol (IP) addresses.
Companies typically configure their directory servers such that each server stores a subset of data types and, notably, the subsets (data types) do not overlap. For instance, a company may have two LDAP servers (Server A and Server B). All corporate human resource related information (employee IDs, email and residential addresses, emergency contacts, salaries, etc) are stored on LDAP Server A, whereas all corporate research and development work, including the various projects under development along with interactions between development groups (both external and internal to the company), are stored on LDAP Server B. Having a database use a plurality of database servers is referred to as a “distributed database”, and a system using a distributed database is referred to as a “distributed database system”.
The subsets of data stored on the LDAP servers are thus reduced and non-overlapping, primarily to avoid overloading each server. LDAP is a database which operates on a schema, i.e., a format of data that the database stores and understands. A directory server (such as LDAP or NDAP) that is configured to increase the amount of data types it stores (e.g., all possible data formats used in an organization) has a complex schema and processing (including searching) of any request is time consuming and inefficient. Attempts by an organization to develop a searching algorithm for such a schema involve use of hash-based, index searching; however, such searching is also quite complex, resulting in overloading of the server and degradation of its performance.
Hash-based indexing is a way of formulating hints that result in faster look-ups; yet indexes generally consume substantial overhead (such as memory and processor cycles) when developing keys for searching the database. Moreover, updates to a hash-based index searching service may adversely affect processing
Ganguly Sukanta
Kuo Chin-Ming
Duong Oanh
Novell Inc.
Schwegman Lundberg Woessner & Kluth P.A.
Sheikh Ayaz
LandOfFree
Predicate indexing of data stored in a computer with... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Predicate indexing of data stored in a computer with..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Predicate indexing of data stored in a computer with... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3071412