Patent application title: BUSINESS INTELLIGENCE MONITORING TOOL
Robert Karch (Greenwich, CT, US)
IPC8 Class: AG06F1730FI
Class name: Record, file, and data search and comparisons database query processing distributed search and retrieval
Publication date: 2016-05-26
Patent application number: 20160147842
An improved method and apparatus for auditing database queries, wherein
comments are added by a server prior to the server forwarding the queries
to the database system. The comments are then used to derive audit
information, which is correlated with other audit information in a
different server, to produce a more complete audit record.
1. A method of logging audit information related to database access, said
method comprising: utilizing a first server to accept database queries
transmitted from plural users to said first server, and logging first
audit information in a first server log, transmitting said queries from
said first server to a database system after logging said first audit
information; capturing second audit information into a second server log
from a communications channel between said first server and a database
system, and constructing an audit record for said queries by correlating
and combining said first audit information and said second audit
2. The method of claim 1 wherein said second audit information includes at least some information that is added to said queries after said queries are transmitted from said plural users to said first server but prior to transmitting said queries from said first server to said database system, and wherein said database system resides on a second server.
3. The method of claim 2 wherein said first audit information includes at least one of report name, universe/catalog name, and user id of a user initiating at least one of said queries through said first server.
4. The method of claim 3 wherein said correlating and combining includes matching a time of transaction recorded in said first server log with a time of transaction recorded in said second server log.
5. The method of claim 4 wherein a user initiating said at least one of said queries does so through a thin client in communication with said first server.
6. The method of claim 2 wherein said audit information is added as a comment.
7. The method of claim 1 wherein said database system is accessed via a Software Query Language (SQL) query.
8. The method of claim 1 wherein the first audit information includes first sets of parameters for respective reports logged in said first server log.
9. The method of claim 8 wherein the second audit information includes second sets of parameters for respective reports stored in said second server log.
10. The method of claim 9 further comprising correlating the first audit information with the second audit information using a parameter in common among the first and second sets of parameters
11. An apparatus for logging information related to database queries, said apparatus comprising a first server that accepts database queries from plural users, a second server that monitors communications between said first server and a database system, said second server recording a first log from information embedded as comments by said first server in messages transmitted from said first server to a database system, and a third server that combines information in said log with information placed in a second log by said first server, in order to construct a third log of desired information relevant to said queries.
12. The apparatus of claim 11 wherein said third log is used to update rules governing acceptable queries to be made in said database.
13. The apparatus of claim 12 wherein said information placed in the second log by said first server includes a user ID to identify a user that initiated a database transaction.
14. A method comprising receiving queries transmitted from plural users, consolidating the queries prior to presenting said queries to a database but after tagging the queries with audit information, and generating an audit file from queries presented to said database, said audit file including said audit information.
15. The method of claim 14 wherein said audit information is user identities.
16. The method of claim 14 further comprising displaying at least some of the information added in said tagging.
17. The method of claim 14 wherein said queries are SQL queries.
CROSS-REFERENCE TO RELATED APPLICATIONS
 This application is a Continuation of U.S. PTO Ser. No. 13/022,699, filed on Feb. 8, 2011, which is a Continuation of U.S. PTO Ser. No. 10/919,643, filed on Aug. 17, 2004, each of which is incorporated herein by reference in their entirety.
 This invention relates to business intelligence (BI) system monitoring tools, and more specifically, to an improved tool for monitoring BI data base queries and related transactions, and for producing a more complete audit trail, then was possible in prior systems.
BACKGROUND OF THE INVENTION
 Business intelligence monitoring tools are typically utilized in connection with large data base access systems in relatively large corporations, as well as in other environments. Typically, such tools monitor and/or control access to databases to increase efficiency, implement security policies, and perform a variety of other functions. As various BI reports are run, information accessed and utilized by a relatively large number of users is often monitored and controlled.
 Many business intelligence systems run on a web server, and permit plural users utilizing, for example, a thin client such as a web browser, to access the database. An exemplary such prior art arrangement is shown in FIG. 1. Plural clients 102-104 connect to the database system 106 through a server with a BI system, shown as 101.
 Many organizations provide monitoring capability shown as 105, in order to monitor the reports and queries taking place between the data base system 106 and the clients through server 101.
 One problem with such prior art arrangement is that it is customary for the server with BI tool 101 to utilize a single database login to process the various requests for reports, queries, and other transactions being communicated between the users 102-104 and the database system 106. More specifically, database 106 and server 101 interface as if server 101 were a single user of the database system invoking various transactions. The server 101 includes appropriate software to then separate out responses from database 106 and forward them to the appropriate one of users 102-104.
 Consequently, queries and other transactions arriving at server 101 from users 102-104 are often stripped of user IDs and other relevant information, and subsequently are sent to database system 106. Accordingly, the communications between database 106 and server 101 includes little or no indication of the specific user 102-104 for which those communications are relevant, and may also be lacking other important information.
 Because the internal information within the server 101 is often unavailable to the monitor 105, and is almost always unavailable in real time, an organization monitoring the communications between 106 and server 101 cannot produce an audit trail that is capable of identifying specific users and other specific information that the server 101 has stripped off.
 In view of the foregoing, prior art systems are only capable of generating somewhat incomplete audit and monitoring data, and a need exists for a more accurate way for organizations to monitor and control database query and access. Additionally, since tools that restrict certain queries based upon database models may rely upon past experience with particular queries, it is important that any monitored information be available for use in updating rules and other systems that may provide security or other access restrictions.
SUMMARY OF THE INVENTION
 The above and other problems of the prior art are overcome in accordance with the present invention, which creates a system that provides a more thorough and flexible technique of generating an audit trail. Monitored information between the web server and the database is supplemented offline by combining it with additional audit information maintained in an internal log by the BI server. The internal log is preferably accessed after the transactions in issue, preferably off line, to ascertain further information about the transactions, including the queries, reports, and other information. Information in the BI web server's internal event log is correlated with information monitored from the communications between the database and the web server, so that a full audit record may be constructed.
 In another embodiment, software "comments" are added by the web server to one or more queries before those queries are passed from the server to the database system. The monitoring tool may be programmed in advance to understand and log the meaning of particular comments, thereby ascertaining information about the query that would otherwise be unavailable.
BRIEF DESCRIPTION OF THE DRAWINGS
 FIG. 1 shows the basic elements of the prior art business information monitoring tool for database access; and
 FIG. 2 shows a block diagram demonstrating conceptually the principals of the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
 The arrangement of FIG. 2 shows the database system 205 with a data source 204, which may be accessed by a plurality of users 201-203, intended in this example to be "thin clients", and 208 which is a "fat client". Clients 201-203, in the exemplary embodiment shown, access database system 205 through a web server 209, which implements various features of a business intelligence system. A "fat client" such as 208, which may be a PC or other computer, may also access database system 205 directly, without interfacing through a web server. Those of ordinary skill in the art understand the difference between a thin client and a fat client, the former having less storage and other processing capability than the latter.
 Although the business intelligence system implemented on web server 209 may be implemented on fat client 208, this is not necessary for the present invention, and either fat client 208 or web server 209 may implement any subset of features of the business intelligence system. A monitor 206 keeps track of information exchanged between data base system 205 and web server 209, and picks off information relevant to construct audit records in an audit log 207.
 The BI event log 210 represents the "internal" event log of the web server or fat client. Although this BI event log 210 is not necessarily physically located within web server 209 or fat client 208, the BI event log is preferably generated by a web server 209 or fat client 208, as opposed to by an independent monitoring or similar tool such as 206.
 Discussing first the example of thin clients 201 through 203, web server 209 accepts various queries and requests for reports, and utilizes a single database login to connect these clients to the database or may consolidate these independent transactions into a single data stream. Database system 205 often has no information concerning the individual user ids or transactions or queries of users 201-203, but instead, treats web server 209 as a single user. It is left up to web server 209, and the BI tool and/or other software therein, to separate out the various queries and other information to and from each of the individual thin clients 201-203.
 In accordance with the invention, the business intelligence tool in web server 209 adds a "comment" to the various database queries transmitted from web server 209 to database system 205. This information added as a comment does not affect the operation of the database system and is ignored by database system 205. Such comments are typically used in software system so that a human reviewing the software can more easily ascertain the purpose of particular software instructions.
 However, in the present invention, the comment added is preferably information that monitor 206 is set up to intercept and monitor. One possibility may be to add a user ID as a comment, so that the data stream between web server 209 and database 205 can be utilized to ascertain the particular user 201-203 that represents the source of the query. Otherwise, such information would be unavailable. Any information desirable may be added as a comment and similarly used, provided that the monitor system knows the intended meaning of the comment, and that the information required to generate the comment is available to the server or fat client to append.
 Monitor 206 may then pick off this information and maintain audit records that include the particular queries, as well as the time they were issued and any other information ascertainable from the communications between web server 209 and database system 205.
 The BI event log 210 is an event log maintained by the web server 209 business intelligence tool, or alternatively, by a business intelligence tool running on a fat client 208. The audit records log may then be generated by examining information monitored by monitor 206, as well as information gathered off line, at a later time during a modeling and analysis process 212, concerning the transaction that's stored in event log 210.
 One example of the combination of information monitored from monitor 206 being combined to create an audit record with information in event log 210 relates to the name of the particular report being run. The web server 209 translates the request for a particular report into a series of software query language (SQL) commands, typical in such database systems. The monitor 206 detects the particular SQL commands being executed and stores them in the Audit Records Log 207. Many BI systems maintain information about reports in a repository 211. Information kept in this repository 211 includes the SQL statement or statements that will be generated by the report.
 The modeling and analysis process 212 will pre-model this report information and store it in its knowledgebase 207. When the SQL is subsequently logged by the monitor 206 into the Audit Records Log 207, the modeling and analysis process 212 will match the pre-modeled SQL to the monitored SQL and thus can obtain knowledge of what particular report those commands correspond to. Hence, the audit records log 207 can contain a record that a particular report was run at a particular time, because the sequence of SQL commands intercepted by monitor 206 commonly corresponds to a single particular report.
 However, in instances where plural reports generate the same SQL sequence, it is difficult for monitor 206 to ascertain which particular report is being run. This problem may be solved by supplementing the information gained by monitor 206 and the modeling process 212 with information stored by the BI event log 210. Wherein the event log 210 may be read off line and correlated with the information in audit records log 207 by the analysis process 212.
 More specifically, BI event log 210 may contain the name of a particular report requested by any one of users 201-203 at a particular time. By comparing the report name and time of execution to the SQL commands captured just after that time by monitor 206 and stored in the audit records log 207, the system can ascertain which particular report was in fact run. Put another way, if the SQL commands captured at a particular time indicate that either report 1 or report 2 were run, and the BI event log 210 indicates that immediately prior to the time those SQL commands were generated, report No. 2 was run, and that report No. 1 was not run at any time close to the time the SQL commands in issue were captured by monitor 206, then report No. 2 must be the one to which the SQL captured by monitor 206 corresponds.
 In a typical exemplary embodiment, three parameters that are included in the audit are the universe/catalog name that a particular report or query accesses, the report name, and the User ID. One or more parameters may be added as a comment by server 209, and one or more of the parameters may be obtained from the BI event log database. The universe/catalogue name point to a set of metadata which describes a set of tables and/or other data that may be derived from such tables and which is relevant to one or more specific reports or database queries.
 There is also a need to correlate the information gathered at monitor 206 with the information relating to that same transaction stored in BI event log 206. This correlation may be done by time of day for example, where the time stamp in audit records log 207 will typically be slightly after the timestamp in BI event log 210, allowing for processing and propagation time through web server 209. Allowances must be made when the BI event log 210 and the Audit Records Log 207 are generated from different physical servers. The clocks on the individual servers may not be syncronized. The modeling and analysis process 212 must utilize a time difference adjustment when matching the timestamps.
 Additionally, various combinations of checksums and serial numbers may be associated with each transaction, in order to correlate the SQL from the BI repository 211 with the data from Audit Records Log 207.
 While the foregoing describes the preferred embodiments of the invention, various other modifications and additions will be apparent to those of skill in the art. The foregoing is intended to be exemplary only, and the invention is defined only by the claims appended hereto.