Patent Grant
8195602
1. Field of Invention
The present invention relates to a method of and system for aggregating data elements in a multi-dimensional database (MDDB) supported upon a computing platform and also to provide an improved method of and system for managing data elements within an MDDB during on-line analytical processing (OLAP) operations and as an integral part of a database management system.
2. Brief Description of the State of the Art
The ability to act quickly and decisively in today's increasingly competitive marketplace is critical to the success of organizations. The volume of information that is available to corporations is rapidly increasing and frequently overwhelming. Those organizations that will effectively and efficiently manage these tremendous volumes of data, and use the information to make business decisions, will realize a significant competitive advantage in the marketplace.
Data warehousing, the creation of an enterprise-wide data store, is the first step towards managing these volumes of data. The Data Warehouse is becoming an integral part of many information delivery systems because it provides a single, central location where a reconciled version of data extracted from a wide variety of operational systems is stored. Over the last few years, improvements in price, performance, scalability, and robustness of open computing systems have made data warehousing a central component of Information Technology CIT strategies. Details on methods of data integration and constructing data warehouses can be found in the white paper entitled “Data Integration: The Warehouse Foundation” by Louis Rolleigh and Joe Thomas.
Building a Data Warehouse has its own special challenges (e.g. using common data model, common business dictionary, etc.) and is a complex endeavor. However, just having a Data Warehouse does not provide organizations with the often-heralded business benefits of data warehousing. To complete the supply chain from transactional systems to decision maker, organizations need to deliver systems that allow knowledge workers to make strategic and tactical decisions based on the information stored in these warehouses. These decision support systems are referred to as On-Line Analytical Processing (OLAP) systems. OLAP systems allow knowledge workers to intuitively, quickly, and flexibly manipulate operational data using familiar business terms, in order to provide analytical insight into a particular problem or line of inquiry. For example, by using an OLAP system, decision makers can “slice and dice” information along a customer (or business) dimension, and view business metrics by product and through time. Reports can be defined from multiple perspectives that provide a high-level or detailed view of the performance of any aspect of the business. Decision makers can navigate throughout their database by drilling down on a report to view elements at finer levels of detail, or by pivoting to view reports from different perspectives. To enable such full-functioned business analyses, OLAP systems need to (1) support sophisticated analyses, (2) scale to large numbers of dimensions, and (3) support analyses against large atomic data sets. These three key requirements are discussed further below.
Decision makers use key performance metrics to evaluate the operations within their domain, and OLAP systems need to be capable of delivering these metrics in a user-customizable format. These metrics may be obtained from the transactional databases pre-calculated and stored in the database, or generated on demand during the query process. Commonly used metrics include:
Knowledge workers analyze data from a number of different business perspectives or dimensions. As used hereinafter, a dimension is any element or hierarchical combination of elements in a data model that can be displayed orthogonally with respect to other combinations of elements in the data model. For example, if a report lists sales by week, promotion, store, and department, then the report would be a slice of data taken from a four-dimensional data model.
Target marketing and market segmentation applications involve extracting highly qualified result sets from large volumes of data. For example, a direct marketing organization might want to generate a targeted mailing list based on dozens of characteristics, including purchase frequency, size of the last purchase, past buying trends, customer location, age of customer, and gender of customer. These applications rapidly increase the dimensionality requirements for analysis.
The number of dimensions in OLAP systems range from a few orthogonal dimensions to hundreds of orthogonal dimensions. Orthogonal dimensions in an exemplary OLAP application might include Geography, Time, and Products.
Atomic data refers to the lowest level of data granularity required for effective decision making. In the case of a retail merchandising manager, “atomic data” may refer to information by store, by day, and by item. For a banker, atomic data may be information by account, by transaction, and by branch. Most organizations implementing OLAP systems find themselves needing systems that can scale to tens, hundreds, and even thousands of gigabytes of atomic information.
As OLAP systems become more pervasive and are used by the majority of the enterprise, more data over longer time frames will be included in the data store (i.e. data warehouse), and the size of the database will increase by at least an order of magnitude. Thus, OLAP systems need to be able to scale from present to near-future volumes of data.
In general, OLAP systems need to (1) support the complex analysis requirements of decision-makers, (2) analyze the data from a number of different perspectives (i.e. business dimensions), and (3) support complex analyses against large input (atomic-level) data sets from a Data Warehouse maintained by the organization using a relational database management system (RDBMS).
Vendors of OLAP systems classify OLAP Systems as either Relational OLAP (ROLAP) or Multidimensional OLAP (MOLAP) based on the underlying architecture thereof. Thus, there are two basic architectures for On-Line Analytical Processing systems: the ROLAP Architecture, and the MOLAP architecture.
The Relational OLAP (ROLAP) system accesses data stored in a Data Warehouse to provide OLAP analyses. The premise of ROLAP is that OLAP capabilities are best provided directly against the relational database, i.e. the Data Warehouse.
The ROLAP architecture was invented to enable direct access of data from Data Warehouses, and therefore support optimization techniques to meet batch window requirements and to provide fast response times. Typically, these optimization techniques include application-level table partitioning, pre-aggregate inferencing, denormalization support, and the joining of multiple fact tables.
A typical prior art ROLAP system has a three-tier or layer client/server architecture. The “database layer” utilizes relational databases for data storage, access, and retrieval processes. The “application logic layer” is the ROLAP engine which executes the multidimensional reports from multiple users. The ROLAP engine integrates with a variety of “presentation layers,” through which users perform OLAP analyses.
After the data model for the data warehouse is defined, data from on-line transaction-processing (OLTP) systems is loaded into the relational database management system (RDBMS). If required by the data model, database routines are run to pre-aggregate the data within the RDBMS. Indices are then created to optimize query access times. End users submit multidimensional analyses to the ROLAP engine, which then dynamically transforms the requests into SQL execution plans. The SQL execution plans are submitted to the relational database for processing, the relational query results are cross-tabulated, and a multidimensional result data set is returned to the end user. ROLAP is a fully dynamic architecture capable of utilizing pre-calculated results when they are available, or dynamically generating results from atomic information when necessary.
Multidimensional OLAP (MOLAP) systems utilize a proprietary multidimensional database (MDDB) to provide OLAP analyses. The MDDB is logically organized as a multidimensional array (typically referred to as a multidimensional cube or hypercube or cube) whose rows/columns each represent a different dimension (i.e., relation). A data value is associated with each combination of dimensions (typically referred to as a “coordinate”). The main premise of this architecture is that data must be stored multidimensionally to be accessed and viewed multidimensionally.
As shown in
Information (i.e. basic data) from a variety of operational systems within an enterprise, comprising the Data Warehouse, is loaded into a prior art multidimensional database (MDDB) through a series of batch routines. The Express™ server by the Oracle Corporation is exemplary of a popular server which can be used to carry out the data loading process in prior art MOLAP systems. As shown in
Offset=Months+Product*(# of_Months)+City*(# of_Months*# of_Products)
During an OLAP session, the response time of a multidimensional query on a prior art MDDB depends on how many cells in the MDDB have to be added “on the fly”. As the number of dimensions in the MDDB increases linearly, the number of the cells in the MDDB increases exponentially. However, it is known that the majority of multidimensional queries deal with summarized high level data. Thus, as shown in
As shown in FIGS. 3C1 and 3C2, the raw data loaded into the MDDB is primarily organized at its lowest dimensional hierarchy, and the results of the pre-aggregations are stored in the neighboring parts of the MDDB.
As shown in FIG. 3C2, along the TIME dimension, weeks are the aggregation results of days, months are the aggregation results of weeks, and quarters are the aggregation results of months. While not shown in the figures, along the GEOGRAPHY dimension, states are the aggregation results of cities, countries are the aggregation results of states, and continents are the aggregation results of countries. By pre-aggregating (i.e. consolidating or compiling) all logical subtotals and totals along all dimensions of the MDDB, it is possible to carry out real-time MOLAP operations using a multidimensional database (MDDB) containing both basic (i.e. atomic) and pre-aggregated data. Once this compilation process has been completed, the MDDB is ready for use. Users request OLAP reports by submitting queries through the OLAP Application interface (e.g. using web-enabled client machines), and the application logic layer responds to the submitted queries by retrieving the stored data from the MDDB for display on the client machine.
Typically, in MDDB systems, the aggregated data is very sparse, tending to explode as the number of dimension grows and dramatically slowing down the retrieval process (as described in the report entitled “Database Explosion: The OLAP Report”, incorporated herein by reference). Quick and on line retrieval of queried data is critical in delivering an on-line response for OLAP queries. Therefore, the data structure of the MDDB, and methods of its storing, indexing and handling are dictated mainly by the need for fast retrieval of massive and sparse data.
Different solutions for this problem are disclosed in the following U.S. patents, each of which is incorporated herein by reference in its entirety:
In all the prior art of OLAP servers, the process of storing, indexing and handling MDDB utilize complex data structures to largely improve the retrieval speed, as part of the querying process, at the cost of slowing down the storing and aggregation. The query-bounded structure, that must support fast retrieval of queries in a restricting environment of high sparcity and multi-hierarchies, is not the optimal one for fast aggregation.
In addition to the aggregation process, the Aggregation, Access and Retrieval module is responsible for all data storage, retrieval and access processes. The Logic module is responsible for the execution of OLAP queries. The Presentation module intermediates between the user and the Logic module and provides an interface through which the end users view and request OLAP analyses. The client/server architecture allows multiple users to simultaneously access the multidimensional database.
In summary, general system requirements of OLAP systems include: (1) supporting sophisticated analysis, (2) scaling to a large number of dimensions, and (3) supporting analysis against large atomic data sets.
MOLAP system architecture is capable of providing analytically sophisticated reports and analysis functionality. However, requirements (2) and (3) fundamentally limit MOLAP's capability, because to be effective and to meet end-user requirements, MOLAP databases need a high degree of aggregation.
By contrast, the ROLAP system architecture allows the construction of systems requiring a low degree of aggregation, but such systems are significantly slower than systems based on MOLAP system architecture principles. The resulting long aggregation times of ROLAP systems impose severe limitations on its volumes and dimensional capabilities.
The graphs plotted in
The large volumes of data and the high dimensionality of certain market segmentation applications are orders of magnitude beyond the limits of current multidimensional databases.
ROLAP is capable of higher data volumes. However, the ROLAP architecture, despite its high volume and dimensionality superiority, suffers from several significant drawbacks as compared to MOLAP:
Thus, improved techniques for data aggregation within MOLAP systems would appear to allow the number of dimensions of and the size of atomic (i.e. basic) data sets in the MDDB to be significantly increased, and thus increase the usage of the MOLAP system architecture.
Also, improved techniques for data aggregation within ROLAP systems would appear to allow for maximized query performance on large data volumes, and to reduce the time of partial aggregations that degrades query response, and thus, generally benefit ROLAP system architectures.
Thus, there is a great need in the art for an improved way of and means for aggregating data elements within a multi-dimensional database (MDDB), while avoiding the shortcomings and drawbacks of prior art systems and methodologies.
Modern operational and informational database systems, as described above, typically use a database management system (DBMS) (such as an RDBMS system, object database system, or object/relational database system) as a repository for storing data and querying the data.
Building a Data Warehouse has its own special challenges (e.g. using common data model, common business dictionary, etc.) and is a complex endeavor. However, just having a Data Warehouse does not provide organizations with the often-heralded business benefits of data warehousing. To complete the supply chain from transactional systems to decision maker, organizations need to deliver systems that allow knowledge workers to make strategic and tactical decisions based on the information stored in these warehouses. These decision support systems are referred to as On-Line Analytical Processing (OLAP) systems. Such OLAP systems are commonly classified as Relational OLAP systems or Multi-Dimensional OLAP systems as described above.
The Relational OLAP (ROLAP) system accesses data stored in a relational database (which is part of the Data Warehouse) to provide OLAP analyses. The premise of ROLAP is that OLAP capabilities are best provided directly against the relational database. The ROLAP architecture was invented to enable direct access of data from Data Warehouses, and therefore support optimization techniques to meet batch window requirements and to provide fast response times. Typically, these optimization techniques include application-level table partitioning, pre-aggregate inferencing, denormalization support, and the joining of multiple fact tables.
As described above, a typical ROLAP system has a three-tier or layer client/server architecture. The “database layer” utilizes relational databases for data storage, access, and retrieval processes. The “application logic layer” is the ROLAP engine which executes the multidimensional reports from multiple users. The ROLAP engine integrates with a variety of “presentation layers,” through which users perform OLAP analyses. After the data model for the data warehouse is defined, data from on-line transaction-processing (OLTP) systems is loaded into the relational database management system (RDBMS). If required by the data model, database routines are run to pre-aggregate the data within the RDBMS. Indices are then created to optimize query access times. End users submit multidimensional analyses to the ROLAP engine, which then dynamically transforms the requests into SQL execution plans. The SQL execution plans are submitted to the relational database for processing, the relational query results are cross-tabulated, and a multidimensional result data set is returned to the end user. ROLAP is a fully dynamic architecture capable of utilizing pre-calculated results when they are available, or dynamically generating results from the raw information when necessary.
The Multidimensional OLAP (MOLAP) systems utilize a proprietary multidimensional database (MDDB) (or “cube”) to provide OLAP analyses. The main premise of this architecture is that data must be stored multidimensionally to be accessed and viewed multidimensionally. Such MOLAP systems provide an interface that enables users to query the MDDB data structure such that users can “slice and dice” the aggregated data. As shown in
There are other application domains where there is a great need for improved methods of and apparatus for carrying out data aggregation operations. For example, modern operational and informational databases represent such domains. As described above, modern operational and informational databases typically utilize a relational database system (RDBMS) as a repository for storing data and querying data.
The choice of using an RDBMS as the data repository in information database systems naturally stems from the realities of SQL standardization, the wealth of RDBMS-related tools, and readily available expertise in RDBMS systems. However, the querying component of RDBMS technology suffers from performance and optimization problems stemming from the very nature of the relational data model. More specifically, during query processing, the relational data model requires a mechanism that locates the raw data elements that match the query. Moreover, to support queries that involve aggregation operations, such aggregation operations must be performed over the raw data elements that match the query. For large multi-dimensional databases, a naive implementation of these operations involves computational intensive table scans that lead to unacceptable query response times.
In order to better understand how the prior art has approached this problem, it will be helpful to briefly describe the relational database model. According to the relational database model, a relational database is represented by a logical schema and tables that implement the schema. The logical schema is represented by a set of templates that define one or more dimensions (entities) and attributes associated with a given dimension. The attributes associated with a given dimension includes one or more attributes that distinguish it from every other dimension in the database (a dimension identifier). Relationships amongst dimensions are formed by joining attributes. The data structure that represents the set of templates and relations of the logical schema is typically referred to as a catalog or dictionary. Note that the logical schema represents the relational organization of the database, but does not hold any fact data per se. This fact data is stored in tables that implement the logical schema.
Star schemas are frequently used to represent the logical structure of a relational database. The basic premise of star schemas is that information can be classified into two groups: facts and dimensions. Facts are the core data elements being analyzed. For example, units of an individual item sold are facts, while dimensions are attributes about the facts. For example, dimensions are the product types purchased and the data purchase. Business questions against this schema are asked by looking up specific facts (UNITS) through a set of dimensions (MARKETS, PRODUCTS, PERIOD). The central fact table is typically much larger than any of its dimension tables.
An exemplary star schema is illustrated in
When processing a query, the tables that implement the schema are accessed to retrieve the facts that match the query. For example, in a star schema implementation as described above, the facts are retrieved from the central fact table and/or the dimension tables. Locating the facts that match a given query involves one or more of the join operations. Moreover, to support queries that involve aggregation operations, such aggregation operations must be performed over the facts that match the query. For large multi-dimensional databases, a naive implementation of these operations involves computational intensive table scans that typically lead to unacceptable query response times. Moreover, since the fact tables are pre-summarized and aggregated along business dimensions, these tables tend to be very large. This point becomes an important consideration of the performance issues associated with star schemas. A more detailed discussion of the performance issues (and proposed approaches that address such issues) related to joining and aggregation of star schema is now set forth.
The first performance issue arises from computationally intensive table scans that are performed by a naive implementation of data joining. Indexing schemes may be used to bypass these scans when performing joining operations. Such schemes include B-tree indexing, inverted list indexing and aggregate indexing. A more detailed description of such indexing schemes can be found in “The Art of Indexing”, Dynamic Information Systems Corporation, October 1999. All of these indexing schemes replace table scan operations (involved in locating the data elements that match a query) with one or more index lookup operation. Inverted list indexing associates an index with a group of data elements, and stores (at a location identified by the index) a group of pointers to the associated data elements. During query processing, in the event that the query matches the index, the pointers stored in the index are used to retrieve the corresponding data elements pointed therefrom. Aggregation indexing integrates an aggregation index with an inverted list index to provide pointers to raw data elements that require aggregation, thereby providing for dynamic summarization of the raw data elements that match the user-submitted query.
These indexing schemes are intended to improve the join operations by replacing table scan operations with one or more index lookup operation in order to locate the data elements that match a query. However, these indexing schemes suffer from various performance issues as follows:
Another performance issue arises from dimension tables that contain multiple hierarchies. In such cases, the dimensional table often includes a level of hierarchy indicator for every record. Every retrieval from the fact table that stores details and aggregates must use the indicator to obtain the correct result, which impacts performance. The best alternative to using the level indicator is the snowflake schema. In this schema, aggregate tables are created separately from the detail tables. In addition to the main fact tables, snowflake schema contains separate fact tables for each level of aggregation. Notably, the snowflake schema is even more complicated than a star schema, and often requires multiple SQL statements to get the results that are required.
Another performance issue arises from the pairwise join problem. Traditional RDBMS engines are not design for the rich set of complex queries that are issued against a star schema. The need to retrieve related information from several tables in a single query—“join processing”—is severely limited. Many RDBMSs can join only two tables at a time. If a complex join involves more than two tables, the RDBMS needs to break the query into a series of pairwise joins. Selecting the order of these joins has a dramatic performance impact. There are optimizers that spend a lot of CPU cycles to find the best order in which to execute those joins. Unfortunately, because the number of combinations to be evaluated grows exponentially with the number of tables being joined, the problem of selecting the best order of pairwise joins rarely can be solved in a reasonable amount of time.
Moreover, because the number of combinations is often too large, optimizers limit the selection on the basis of a criterion of directly related tables. In a star schema, the fact table is the only table directly related to most other tables, meaning that the fact table is a natural candidate for the first pairwise join. Unfortunately, the fact table is the very largest table in the query, so this strategy leads to selecting a pairwise join order that generates a very large intermediate result set, severely affecting query performance.
There is an optimization strategy, typically referred to as Cartesian Joins, that lessens the performance impact of the pairwise join problem by allowing joining of unrelated tables. The join to the fact table, which is the largest one, is deferred until the very end, thus reducing the size of intermediate result sets. In a join of two unrelated tables every combination of the two tables' rows is produced, a Cartesian product. Such a Cartesian product improves query performance. However, this strategy is viable only if the Cartesian product of dimension rows selected is much smaller than the number of rows in the fact table. The multiplicative nature of the Cartesian join makes the optimization helpful only for relatively small databases.
In addition, systems that exploit hardware and software parallelism have been developed which lessens the performance issues set forth above. Parallelism can help reduce the execution time of a single query (speed-up), or handle additional work without degrading execution time (scale-up). For example, Red Brick™ has developed STARjoin™ technology that provides high speed, parallelizable multi-table joins in a single pass, thus allowing more than two tables to be joined in a single operation. The core technology is an innovative approach to indexing that accelerates multiple joins. Unfortunately, parallelism can only reduce, not eliminate, the performance degradation issues related to the star schema.
One of the most fundamental principles of the multidimensional database is the idea of aggregation. The most common aggregation is called a roll-up aggregation. This type is relatively easy to compute: e.g. taking daily sales totals and rolling them up into a monthly sales table. The more difficult are analytical calculations, the aggregation of Boolean and comparative operators. However these are also considered as a subset of aggregation.
In a star schema, the results of aggregation are summary tables. Typically, summary tables are generated by database administrators who attempt to anticipate the data aggregations that the users will request, and then pre-build such tables. In such systems, when processing a user-generated query that involves aggregation operations, the pre-built aggregated data that matches the query is retrieved from the summary tables (if such data exists).
Summary tables containing pre-aggregated results typically provide for improved query response time with respect to on-the-fly aggregation. However, summary tables suffer from some disadvantages:
Note that in the event that the aggregated data does not exist in the summary tables, the table join operations and the aggregation operations are performed over the raw facts in order to generate such aggregated data. This is typically referred to as on-the-fly aggregation. In such instances, aggregation indexing is used to mitigate the performance of multiple data joins associated with dynamic aggregation of the raw data. Thus, in large multi-dimensional databases, such dynamic aggregation may lead to unacceptable query response times.
In view of the problems associated with joining and aggregation within RDBMS, prior art ROLAP systems have suffered from essentially the same shortcomings and drawbacks of their underlying RDBMS.
While prior art MOLAP systems provide for improved access time to aggregated data within their underlying MDD structures, and have performance advantages when carrying out joining and aggregation operations, prior art MOLAP architectures have suffered from a number of shortcomings and drawbacks. More specifically, atomic (raw) data is moved, in a single transfer, to the MOLAP system for aggregation, analysis and querying. Importantly, the aggregation results are external to the DBMS. Thus, users of the DBMS cannot directly view these results. Such results are accessible only from the MOLAP system. Because the MDD query processing logic in prior art MOLAP systems is separate from that of the DBMS, users must procure rights for access to the MOLAP system and be instructed (and be careful to conform to such instructions) to access the MDD (or the DBMS) under certain conditions. Such requirements can present security issues, highly undesirable for system administration. Satisfying such requirements is a costly and logistically cumbersome process. As a result, the widespread applicability of MOLAP systems has been limited.
Thus, there is a great need in the art for an improved mechanism for joining and aggregating data elements within a database management system (e.g., RDBMS), and for integrating the improved database management system (e.g., RDBMS) into informational database systems (including the data warehouse and OLAP domains), while avoiding the shortcomings and drawbacks of prior art systems and methodologies.
Accordingly, it is a further object of the present invention to provide an improved method of and system for managing data elements within a multidimensional database (MDDB) using a novel stand-alone (i.e. external) data aggregation server, achieving a significant increase in system performance (e.g. decreased access/search time) using a stand-alone scalable data aggregation server.
Another object of the present invention is to provide such a system, wherein the stand-alone aggregation server includes an aggregation engine that is integrated with an MDDB, to provide a cartridge-style plug-in accelerator which can communicate with virtually any conventional OLAP server.
Another object of the present invention is to provide such a stand-alone data aggregration server whose computational tasks are restricted to data aggregation, leaving all other OLAP functions to the MOLAP server and therefore complementing OLAP server's functionality.
Another object of the present invention is to provide such a system, wherein the stand-alone aggregation server carries out an improved method of data aggregation within the MDDB which enables the dimensions of the MDDB to be scaled up to large numbers and large atomic (i.e. base) data sets to be handled within the MDDB.
Another object of the present invention is to provide such a stand-alone aggregration server, wherein the aggregation engine supports high-performance aggregation (i.e. data roll-up) processes to maximize query performance of large data volumes, and to reduce the time of partial aggregations that degrades the query response.
Another object of the present invention is to provide such a stand-alone, external scalable aggregation server, wherein its integrated data aggregation (i.e. roll-up) engine speeds up the aggregation process by orders of magnitude, enabling larger database analysis by lowering the aggregation times.
Another object of the present invention is to provide such a novel stand-alone scalable aggregation server for use in OLAP operations, wherein the scalability of the aggregation server enables (i) the speed of the aggregation process carried out therewithin to be substantially increased by distributing the computationally intensive tasks associated with data aggregation among multiple processors, and (ii) the large data sets contained within the MDDB of the aggregation server to be subdivided among multiple processors thus allowing the size of atomic (i.e. basic) data sets within the MDDB to be substantially increased.
Another object of the present invention is to provide such a novel stand-alone scalable aggregation server, which provides for uniform load balancing among processors for high efficiency and best performance, and linear scalability for extending the limits by adding processors.
Another object of the present invention is to provide a stand-alone, external scalable aggregation server, which is suitable for MOLAP as well as for ROLAP system architectures.
Another object of the present invention is to provide a novel stand-alone scalable aggregation server, wherein an MDDB and aggregation engine are integrated and the aggregation engine carries out a high-performance aggregation algorithm and novel storing and searching methods within the MDDB.
Another object of the present invention is to provide a novel stand-alone scalable aggregation server which can be supported on single-processor (i.e. sequential or serial) computing platforms, as well as on multi-processor (i.e. parallel) computing platforms.
Another object of the present invention is to provide a novel stand-alone scalable aggregation server which can be used as a complementary aggregation plug-in to existing MOLAP and ROLAP databases.
Another object of the present invention is to provide a novel stand-alone scalable aggregation server which carries out novel rollup (i.e. down-up) and spread down (i.e. top-down) aggregation algorithms.
Another object of the present invention is to provide a novel stand-alone scalable aggregation server which includes an integrated MDDB and aggregation engine which carries out full pre-aggregation and/or “on-the-fly” aggregation processes within the MDDB.
Another object of the present invention is to provide such a novel stand-alone scalable aggregation server which is capable of supporting MDDB having a multi-hierarchy dimensionality.
Another object of the present invention is to provide a novel method of aggregating multidimensional data of atomic data sets originating from an RDBMS Data Warehouse.
Another object of the present invention is to provide a novel method of aggregating multidimensional data of atomic data sets originating from other sources, such as external ASCII files, an MOLAP server, or other end user applications.
Another object of the present invention is to provide a novel stand-alone scalable data aggregation server which can communicate with any MOLAP server via standard ODBC, OLE DB or DLL interface, in a completely transparent manner with respect to the (client) user, without any time delays in queries, equivalent to storage in the MOLAP server's cache.
Another object of the present invention is to provide a novel “cartridge-style” (stand-alone) scalable data aggregation engine which dramatically expands the boundaries of MOLAP into large-scale applications including Banking, Insurance, Retail and Promotion Analysis.
Another object of the present invention is to provide a novel “cartridge-style” (stand-alone) scalable data aggregation engine which dramatically expands the boundaries of high-volatility type ROLAP applications such as, for example, the precalculation of data to maximize query performance.
Another object of the present invention is to provide a generic plug-in cartridge-type data aggregation component, suitable for all MOLAP systems of different vendors, dramatically reducing their aggregation burdens.
Another object of the present invention is to provide a novel high performance cartridge-type data aggregration server which, having standardized interfaces, can be plugged-into the OLAP system of virtually any user or vendor.
Another object of the present invention is to provide a novel “cartridge-style” (stand-alone) scalable data aggregation engine which has the capacity to convert long batch-type data aggregations into interactive sessions.
In another aspect, it is an object of the present invention to provide an improved method of and system for joining and aggregating data elements integrated within a database management system (DBMS) using a non-relational multi-dimensional data structure (MDDB), achieving a significant increase in system performance (e.g. decreased access/search time), user flexibility and ease of use.
Another object of the present invention is to provide such a DBMS wherein its integrated data aggregation module supports high-performance aggregation (i.e. data roll-up) processes to maximize query performance of large data volumes.
Another object of the present invention is to provide such a DBMS system, wherein its integrated data aggregation (i.e. roll-up) module speeds up the aggregation process by orders of magnitude, enabling larger database analysis by lowering the aggregation times.
Another object of the present invention is to provide such a novel DBMS system for use in OLAP operations.
Another object of the present invention is to provide a novel DBMS system having an integrated aggregation module that carries out novel rollup (i.e. down-up) and spread down (i.e. top-down) aggregation algorithms.
Another object of the present invention is to provide a novel DBMS system having an integrated aggregation module that carries out full pre-aggregation and/or “on-the-fly” aggregation processes.
Another object of the present invention is to provide a novel DBMS system having an integrated aggregation module which is capable of supporting an MDDB having a multi-hierarchy dimensionality.
These and other objects of the present invention will become apparent hereinafter and in the Claims to Invention set forth herein.
In order to more fully appreciate the objects of the present invention, the following Detailed Description of the Illustrative Embodiments should be read in conjunction with the accompanying Drawings, wherein:
FIG. 3C1 is a schematic representation of an exemplary three-dimensional database used in a conventional MOLAP system of the prior art, showing that each data element contained therein is physically stored at a location in the recording media of the system which is specified by the dimensions (and subdimensions within the dimensional hierarchy) of the data variables which are assigned integer-based coordinates in the MDDB, and also that data elements associated with the basic data loaded into the MDDB are assigned lower integer coordinates in MDDB Space than pre-aggregated data elements contained therewithin;
FIG. 3C2 is a schematic representation illustrating that a conventional hierarchy of the dimension of “time” typically contains the subdimensions “days, weeks, months, quarters, etc.” of the prior art;
FIG. 3C3 is a schematic representation showing how data elements having higher subdimensions of time in the MDDB of the prior art are typically assigned increased integer addresses along the time dimension thereof;
FIG. 9C1 is a schematic representation of the Query Directed Roll-up (QDR) aggregation method/procedure of the present invention, showing data aggregation starting from existing basic data or previously aggregated data in the first dimension (D1), and such aggregated data being utilized as a basis for QDR aggregation along the second dimension (D2);
FIG. 9C2 is a schematic representation of the Query Directed Roll-up (QDR) aggregation method/procedure of the present invention, showing initial data aggregation starting from existing previously aggregated data in the third dimension (D3), and continuing along the third dimension (D3), and thereafter continuing aggregation along the second dimension (D2);
FIG. 11C(i) through 11C(ix) represent a flow chart description (and accompanying data structures) of the operations of an exemplary hierarchy transformation mechanism of the present invention that optimally merges multiple hierarchies into a single hierarchy that is functionally equivalent to the multiple hierarchies.
FIGS. 19C(i) and 19C(ii), taken together, set forth a flow chart representation of the primary operations carried out within the DBMS of the present invention when performing data aggregation and related support operations, including the servicing of user-submitted (e.g. natural language) queries made on such aggregated database of the present invention.
Referring now to
Through this invention disclosure, the term “aggregation” and “preaggregation” shall be understood to mean the process of summation of numbers, as well as other mathematical operations, such as multiplication, subtraction, division etc.
In general, the stand-alone aggregation server and methods of and apparatus for data aggregation of the present invention can be employed in a wide range of applications, including MOLAP systems, ROLAP systems, Internet URL-directory systems, personalized on-line e-commerce shopping systems, Internet-based systems requiring real-time control of packet routing and/or switching, and the like.
For purposes of illustration, initial focus will be accorded to improvements in MOLAP systems, in which knowledge workers are enabled to intuitively, quickly, and flexibly manipulate operational data within a MDDB using familiar business terms in order to provide analytical insight into a business domain of interest.
Departing from conventional practices, the principles of the present invention teaches moving the aggregation engine and the MDDB into a separate Aggregation Server having standardized interfaces so that it can be plugged-into the OLAP server of virtually any user or vendor. This dramatic move discontinues the restricting dependency of aggregation from the analytical functions of OLAP, and by applying novel and independent algorithms. The stand-alone data aggregation server enables efficient organization and handling of data, fast aggregation processing, and fast access to and retrieval of any data element in the MDDB.
As will be described in greater detail hereinafter, the Aggregation Server of the present invention can serve the data aggregation requirements of other types of systems besides OLAP systems such as, for example, URL directory management Data Marts, RDBMS, or ROLAP systems.
The Aggregation Server of the present invention excels in performing two distinct functions, namely: the aggregation of data in the MDDB; and the handling of the resulting data base in the MDDB, for “on demand” client use. In the case of serving an OLAP server, the Aggregation Server of the present invention focuses on performing these two functions in a high performance manner (i.e. aggregating and storing base data, originated at the Data Warehouse, in a multidimensional storage (MDDB), and providing the results of this data aggregation process “on demand” to the clients, such as the OLAP server, spreadsheet applications, the end user applications. As such, the Aggregation Server of the present invention frees each conventional OLAP server, with which it interfaces, from the need of making data aggregations, and therefore allows the conventional OLAP server to concentrate on the primary functions of OLAP servers, namely: data analysis and supporting a graphical interface with the user client.
During operation, the base data originates at data warehouse or other sources, such as external ASCII files, MOLAP server, or others. The Configuration Manager, in order to enable proper communication with all possible sources and data structures, configures two blocks, the Base Data Interface and Data Loader. Their configuration is matched with different standards such as OLDB, OLE-DB, ODBC, SQL, API, JDBC, etc.
As shown in
As shown in
An object of the present invention is to make the transfer of data completely transparent to the OLAP user, in a manner which is equivalent to the storing of data in the MOLAP server's cache and without any query delays. This requires that the stand-alone Aggregation Server have exceptionally fast response characteristics. This object is enabled by providing the unique data structure and aggregation mechanism of the present invention.
The function of the aggregation management module is to administrate the aggregation process according to the method illustrated in
In accordance with the principles of the present invention, data aggregation within the stand-alone Aggregation Server can be carried out either as a complete pre-aggregation process, where the base data is fully aggregated before commencing querying, or as a query directed roll-up (QDR) process, where querying is allowed at any stage of aggregation using the “on-the-fly” data aggregation process of the present invention. The QDR process will be described hereinafter in greater detail with reference to
The function of the Storage management module is to handle multidimensional data in the storage(s) module in a very efficient way, according to the novel method of the present invention, which will be described in detail hereinafter with reference to
The request serving mechanism shown in
The segmented data aggregation method of the present invention is described in FIGS. 9A through 9C2. These figures outline a simplified setting of three dimensions only; however, the following analysis applies to any number of dimensions as well.
The data is being divided into autonomic segments to minimize the amount of simultaneously handled data. The initial aggregation is practiced on a single dimension only, while later on the aggregation process involves all other dimensions.
At the first stage of the aggregation method, an aggregation is performed along dimension 1. The first stage can be performed on more than one dimension. As shown in
In the next stage shown in
The principle of data segmentation can be applied on the first stage as well. However, only a large enough data set will justify such a sliced procedure in the first dimension. Actually, it is possible to consider each segment as an N−1 cube, enabling recursive computation.
It is imperative to get aggregation results of a specific slice before the entire aggregation is completed, or alternatively, to have the roll-up done in a particular sequence. This novel feature of the aggregation method of the present invention is that it allows the querying to begin, even before the regular aggregation process is accomplished, and still having fast response. Moreover, in relational OLAP and other systems requiring only partial aggregations, the QDR process dramatically speeds up the query response.
The QDR process is made feasible by the slice-oriented roll-up method of the present invention. After aggregating the first dimension(s), the multidimensional space is composed of independent multidimensional cubes (slices). These cubes can be processed in any arbitrary sequence.
Consequently the aggregation process of the present invention can be monitored by means of files, shared memory sockets, or queues to statically or dynamically set the roll-up order.
In order to satisfy a single query coming from a client, before the required aggregation result has been prepared, the QDR process of the present invention involves performing a fast on-the-fly aggregation (roll-up) involving only a thin slice of the multidimensional data.
FIG. 9C1 shows a slice required for building-up a roll-up result of the 2nd dimension. In case 1, as shown, the aggregation starts from an existing data, either basic or previously aggregated in the first dimension. This data is utilized as a basis for QDR aggregation along the second dimension. In case 2, due to lack of previous data, a QDR involves an initial slice aggregation along dimension 3, and thereafter aggregation along the 2nd dimension.
FIG. 9C2 shows two corresponding QDR cases for gaining results in the 3d dimension. Cases 1 and 2 differ in the amount of initial aggregation required in 2nd dimension.
A search for a queried data point is then performed by an access to the DIR file. The search along the file can be made using a simple binary search due to file's ascending order. When the record is found, it is then loaded into main memory to search for the required point, characterized by its index INDk. The attached Data field represents the queried value. In case the exact index is not found, it means that the point is a NA.
In another aspect of the present invention, a novel method is provided for optimally merging multiple hierarchies in multi-hierarchical structures. The method, illustrated in
According to the devised method, the inner order of hierarchies within a dimension is optimized, to achieve efficient data handling for summations and other mathematical formulas (termed in general “Aggregation”). The order of hierarchy is defined externally. It is brought from a data source to the stand-alone aggregation engine, as a descriptor of data, before the data itself. In the illustrative embodiment, the method assumes hierarchical relations of the data, as shown in
Notably, when using prior art techniques, multiple handling of data elements, which occurs when a data element is accessed more than once during aggregation process, has been hitherto unavoidable when the main concern is to effectively handle the sparse data. The data structures used in prior art data handling methods have been designed for fast access to a non NA data. According to prior art techniques, each access is associated with a timely search and retrieval in the data structure. For the massive amount of data typically accessed from a Data Warehouse in an OLAP application, such multiple handling of data elements has significantly degraded the efficiency of prior art data aggregation processes. When using prior art data handling techniques, the data element D shown in
In accordance with the data handling method of the present invention, the data is being pre-ordered for a singular handling, as opposed to multiple handling taught by prior art methods. According to the present invention, elements of base data and their aggregated results are contiguously stored in a way that each element will be accessed only once. This particular order allows a forward-only handling, never backward. Once a base data element is stored, or aggregated result is generated and stored, it is never to be retrieved again for further aggregation. As a result the storage access is minimized. This way of singular handling greatly elevates the aggregation efficiency of large data bases. An efficient handling method as used in the present invention, is shown in
FIG. 11C(i) through 11C(ix) represent a flow chart description (and accompanying data structures) of the operations of an exemplary hierarchy transformation mechanism of the present invention that optimally merges multiple hierarchies into a single hierarchy that is functionally equivalent to the multiple hierarchies. For the sake of description, the data structures correspond to exemplary hierarchical structures described above with respect to
In the loop 1105-1119, a given item in the multiple hierarchy is selected (step 1107); and, in step 1109, the parent(s) (if any)—including grandparents, great-grandparents, etc.—of the given item are identified and added to an entry (for the given item) in a parent list data structure, which is illustrated in FIG. 11C(v). Each entry in the parent list corresponds to a specific item and includes zero or more identifiers for items that are parents (or grandparents, or great-grandparents) of the specific item. In addition, an inner loop (steps 1111-1117) is performed over the hierarchies of the multiple hierarchies described by the hierarchy descriptor data, wherein in step 1113 one of the multiple hierarchies is selected. In step 1115, the child of the given item in the selected hierarchy (if any) is identified and added (if need be) to a group of identifiers in an entry (for the given item) in a child list data structure, which is illustrated in FIG. 11C(vi). Each entry in the child list corresponds to a specific item and includes zero or more groups of identifiers each identifying a child of the specific item. Each group corresponds to one or more of the hierarchies described by the hierarchy descriptor data.
The operation then continues to steps 1121 and 1123 as illustrated in FIG. 11C(ii) to verify the integrity of the multiple hierarchies described by the hierarchy descriptor data (step 1121) and fix (or report to the user) any errors discovered therein (step 1123). Preferably, the integrity of the multiple hierarchies is verified in step 1121 by iteratively expanding each group of identifiers in the child list to include the children, grandchildren, etc of any item listed in the group. If the child(ren) for each group for a specific item do not match, a verification error is encountered, and such error is fixed (or reported to the user (step 1123). The operation then proceeds to a loop (steps 1125-1133) over the items in the child list.
In the loop (steps 1125-1133), a given item in the child list is identified in step 1127. In step 1129, the entry in the child list for the given item is examined to determine if the given item has no children (e.g., the corresponding entry is null). If so, the operation continues to step 1131 to add an entry for the item in level 0 of an ordered list data structure, which is illustrated in FIG. 11C(vii); otherwise the operation continues to process the next item of the child list in the loop. Each entry in a given level of the order list corresponds to a specific item and includes zero or more identifiers each identifying a child of the specific item. The levels of the order list described the transformed hierarchy as will readily become apparent in light of the following. Essentially, loop 1125-1333 builds the lowest level (level 0) of the transformed hierarchy.
After loop 1125-1133, operation continues to process the lowest level to derive the next higher level, and iterate over this process to build out the entire transformed hierarchy. More specifically, in step 1135, a “current level” variable is set to identify the lowest level. In step 1137, the items of the “current level” of the ordered list are copied to a work list. In step 1139, it is determined if the worklist is empty. If so, the operation ends; otherwise operation continues to step 1141 wherein a loop (steps 1141-1159) is performed over the items in the work list.
In step 1143, a given item in the work list is identified and operation continues to an inner loop (steps 1145-1155) over the parent(s) of the given item (which are specified in the parent list entry for the given item). In step 1147 of the inner loop, a given parent of the given item is identified. In step 1149, it is determined whether any other parent (e.g., a parent other than the given patent) of the given item is a child of the given parent (as specified in the child list entry for the given parent). If so, operation continues to step 1155 to process the next parent of the given item in the inner loop; otherwise, operation continues to steps 1151 and 1153. In step 1151, an entry for the given parent is added to the next level (current level +1) of the ordered list, if it does not exist there already. In step 1153, if no children of the given item (as specified in the entry for the given item in the current level of the ordered list) matches (e.g., is covered by) any child (or grandchild or great grandchild etc) of item(s) in the entry for the given parent in the next level of the ordered list, the given item is added to the entry for the given parent in the next level of the ordered list. Levels 1 and 2 of the ordered list for the example described above are shown in FIGS. 11C(viii) and 11C(ix), respectively. The children (including grandchildren and great grandchildren. etc) of an item in the entry for a given parent in the next level of the ordered list may be identified by the information encoded in the lower levels of the ordered list. After step 1153, operation continues to step 1155 to process the next parent of the given item in the inner loop (steps 1145-1155)
After processing the inner loop (steps 1145-1155), operation continues to step 1157 to delete the given item from the work list, and processing continues to step 1159 to process the next item of the work list in the loop (steps 1141-1159).
After processing the loop (steps 1141-1159), the ordered list (e.g., transformed hierarchy) has been built for the next higher level. The operation continues to step 1161 to increment the current level to the next higher level, and operation returns (in step 1163) to step 1138 to build the next higher level, until the highest level is reached (determined in step 1139) and the operation ends.
The reason for the central multidimensional database's rise to corporate necessity is that it facilitates flexible, high-performance access and analysis of large volumes of complex and interrelated data.
A stand-alone specialized aggregation server, simultaneously serving many different kinds of clients (e.g. data mart, OLAP, URL, RDBMS), has the power of delivering an enterprise-wide aggregation in a cost-effective way. This kind of server eliminates the roll-up redundancy over the group of clients, delivering scalability and flexibility.
Performance associated with central data warehouse is an important consideration in the overall approach. Performance includes aggregation times and query response.
Effective interactive query applications require near real-time performance, measured in seconds. These application performances translate directly into the aggregation requirements.
In the prior art, in case of MOLAP, a full pre-aggregation must be done before starting querying. In the present invention, in contrast to prior art, the query directed roll-up (QDR) allows instant querying, while the full pre-aggregation is done in the background. In cases a full pre-aggregation is preferred, the currently invented aggregation outperforms any prior art. For the ROLAP and RDBMS clients, partial aggregations maximize query performance. In both cases fast aggregation process is imperative. The aggregation performance of the current invention is by orders of magnitude higher than that of the prior art.
The stand-alone scalable aggregation server of the present invention can be used in any MOLAP system environment for answering questions about corporate performance in a particular market, economic trends, consumer behaviors, weather conditions, population trends, or the state of any physical, social, biological or other system or phenomenon on which different types or categories of information, organizable in accordance with a predetermined dimensional hierarchy, are collected and stored within a RDBMS of one sort or another. Regardless of the particular application selected, the address data mapping processes of the present invention will provide a quick and efficient way of managing a MDDB and also enabling decision support capabilities utilizing the same in diverse application environments.
The stand-alone “cartridge-style” plug-in features of the data aggregation server of the present invention, provides freedom in designing an optimized multidimensional data structure and handling method for aggregation, provides freedom in designing a generic aggregation server matching all OLAP vendors, and enables enterprise-wide centralized aggregation.
The method of Segmented Aggregation employed in the aggregation server of the present invention provides flexibility, scalability, a condition for Query Directed Aggregation, and speed improvement.
The method of Multidimensional data organization and indexing employed in the aggregation server of the present invention provides fast storage and retrieval, a condition for Segmented Aggregation, improves the storing, handling, and retrieval of data in a fast manner, and contributes to structural flexibility to allow sliced aggregation and QDR. It also enables the forwarding and single handling of data with improvements in speed performance.
The method of Query Directed Aggregation (QDR) employed in the aggregation server of the present invention minimizes the data handling operations in multi-hierarchy data structures.
The method of Query Directed Aggregation (QDR) employed in the aggregation server of the present invention eliminates the need to wait for full aggregation to be completed, and provides build-up aggregated data required for full aggregation.
In another aspect of the present invention, an improved DBMS system (e.g., RDBMS system, object oriented database system or object/relational database system) is provided that excels in performing two distinct functions, namely: the aggregation of data; and the handling of the resulting data for “on demand” client use. Moreover, because of improved data aggregation capabilities, the DBMS of the present invention can be employed in a wide range of applications, including Data Warehouses supporting OLAP systems and the like. For purposes of illustration, initial focus will be accorded to the DBMS of the present invention. Referring now to
Through this document, the term “aggregation” and “pre-aggregation” shall be understood to mean the process of summation of numbers, as well as other mathematical operations, such as multiplication, subtraction, division etc. It shall be understood that pre-aggregation operations occur asynchronously with respect to the traditional query processing operations. Moreover, the term “atomic data” shall be understood to refer to the lowest level of data granularity required for effective decision making. In the case of a retail merchandising manager, atomic data may refer to information by store, by day, and by item. For a banker, atomic data may be information by account, by transaction, and by branch.
It should be noted that the DBMS typically includes additional components (not shown) that are not relevant to the present invention. The query interface and query handler service user-submitted queries (in the preferred embodiment, SQL query statements) forwarded, for example, from a client machine over a network as shown. The query handler and relational data store (tables and meta-data store) are operably coupled to the MDD Aggregation Module. Importantly, the query handler and integrated MDD Aggregation Module operate to provide for dramatically improved query response times for data aggregation operations and drill-downs. Moreover, it is an object of the present invention to make user-querying of the non-relational MDDB no different than querying a relational table of the DBMS, in a manner that minimizes the delays associated with queries that involve aggregation or drill down operations. This object is enabled by providing the novel DBMS system and integrated aggregation mechanism of the present invention.
During operation, base data originates from the table(s) of the DBMS. The core data aggregation operations are performed by the Aggregation Engine; a Multidimensional Data (MDDB) Handler; and a Multidimensional Data Storage (MDDB). The results of data aggregation are efficiently stored in the MDDB by the MDDB Handler. The SQL handler of the MDD Aggregation module services user-submitted queries (in the preferred embodiment, SQL query statements) forwarded from the query handler of the DBMS. The SQL handler of the MDD Aggregation module may communicate with the query handler of the DBMS over a standard interface (such as OLDB, OLE-DB, ODBC, SQL, API, JDBC, etc.). In this case, the support mechanisms of the RDBMS and SQL handler include components that provide communication of such data over these standard interfaces. Such interface components are well known in the art. Aggregation (or drill down results) are retrieved on demand and returned to the user.
Typically, a user interacts with a client machine (for example, using a web-enabled browser) to generate a natural language query, that is communicated to the query interface of the DBMS, for example over a network as shown. The query interface disintegrates the query, via parsing, into a series of requests (in the preferred embodiment, SQL statements) that are communicated to the query handler of the DBMS. It should be noted that the functions of the query interface may be implemented in a module that is not part of the DBMS (for example, in the client machine). The query handler of the DBMS forwards requests that involve data stored in the MDD of the MDD Aggregation module to the SQL hander of the MDD Aggregation module for servicing. Each request specifies a set of n-dimensions. The SQL handler of the MDD Aggregation Module extracts this set of dimensions and operates cooperatively with the MDD handler to address the MDDB using the set of dimensions, retrieve the addressed data from the MDDB, and return the results to the user via the query handler of the DBMS.
FIGS. 19C(i) and 19C(ii) is a flow chart illustrating the operations of an illustrative DBMS of the present invention. In step 601, the base data loader of the MDD Aggregation Module loads the dictionary (or catalog) from the meta-data store of the DBMS. In performing this function, the base data loader may utilize an adapter (interface) that maps the data types of the dictionary of the DBMS (or that maps a standard data type used to represent the dictionary of the DBMS) into the data types used in the MDD aggregation module. In addition, the base data loader extracts the dimensions from the dictionary and forwards the dimensions to the aggregation engine of the MDD Aggregation Module.
In step 603, the base data loader loads table(s) from the DBMS. In performing this function, the base data loader may utilize an adapter (interface) that maps the data types of the table(s) of the DBMS (or that maps a standard data type used to represent the fact table(s) of the DBMS) into the data types used in the MDD Aggregation Module. In addition, the base data loader extracts the atomic data from the table(s), and forwards the atomic data to the aggregation engine.
In step 605, the aggregation engine performs aggregation operations (i.e., roll-up operation) on the atomic data (provided by the base data loader in step 603) along at least one of the dimensions (extracted from the dictionary of the DBMS in step 601) and operates cooperatively with the MDD handler to store the resultant aggregated data in the MDDB. A more detailed description of exemplary aggregation operations according to a preferred embodiment of the present invention is set forth below with respect to the QDR process of
In step 607, a reference is defined that provides users with the ability to query the data generated by the MDD Aggregation Module and/or stored in the MDDB of the MDD Aggregation Module. This reference is preferably defined using the Create View SQL statement, which allows the user to: i) define a table name (TN) associated with the MDDB stored in the MDD Aggregation Module, and ii) define a link used to route SQL statements on the table TN to the MDD Aggregation Module. In this embodiment, the view mechanism of the DBMS enables reference and linking to the data stored in the MDDB of the MDD Aggregation Engine as illustrated in
In step 609, a user interacts with a client machine to generate a query, and the query is communicated to the query interface. The query interface generate one or more SQL statements. These SQL statements may refer to data stored in tables of the relational datastore, or may refer to the reference defined in step 607 (this reference refers to the data stored in the MDDB of the MDD Aggregation Module). These SQL statement(s) are forwarded to the query handler of the DBMS.
In step 611, the query handler receives the SQL statement(s); and optionally transforms such SQL statement(s) to optimize the SQL statement(s) for more efficient query handling. Such transformations are well known in the art. For example, see Kimball, “Aggregation Navigation With (Almost) No MetaData”, DBMS Data Warehouse Supplement, August 1996.
In step 613: the query handler determines whether the received SQL statement(s) [or transformed SQL statement(s)] is on the reference generated in step 607. If so, operation continues to step 615; otherwise normal query handling operations continue in step 625 wherein the relational datastore is accessed to extract, store, and/or manipulate the data stored therein as directed by the query, and results are returned back to the user via the client machine, if needed.
In step 615, the received SQL statement(s) [or transformed SQL statement(s)] is routed to the MDD aggregation engine for processing in step 617 using the link for the reference as described above with respect to step 607.
In step 617, the SQL statement(s) is received by the SQL handler of the MDD Aggregation Module, wherein a set of one or more N-dimensional coordinates are extracted from the SQL statement. In performing this function, SQL handler may utilize an adapter (interface) that maps the data types of the SQL statement issued by query handler of the DBMS (or that maps a standard data type used to represent the SQL statement issued by query handler of the DBMS) into the data types used in the MDD aggregation module.
In step 619, the set of N-dimensional coordinates extracted in step 617 are used by the MDD handler to address the MDDB and retrieve the corresponding data from the MDDB.
Finally, in step 621, the retrieved data is returned to the user via the DBMS (for example, by forwarding the retrieved data to the SQL handler, which returns the retrieved data to the query handler of the DBMS system, which returns the results of the user-submitted query to the user via the client machine), and the operation ends.
It should be noted that the table data (base data), as it arrives from DBMS, may be analyzed and reordered to optimize hierarchy handling, according to the unique method of the present invention, as described above with reference to
Moreover, the MDD control module of the MDD Aggregation Module preferably administers the aggregation process according to the method illustrated in
The SQL handling mechanism shown in
As illustrated in
Functional Advantages Gained by the Improved DBMS of the Present Invention
The features of the DBMS of the present invention, provides for dramatically improved response time in handling queries issued to the DBMS that involve aggregation, thus enabling enterprise-wide centralized aggregation. Moreover, in the preferred embodiment of the present invention, users can query the aggregated data in an manner no different than traditional queries on the DBMS.
The method of Segmented Aggregation employed by the novel DBMS of the present invention provides flexibility, scalability, the capability of Query Directed Aggregation, and speed improvement.
Moreover, the method of Query Directed Aggregation (QDR) employed by the novel DBMS of the present invention minimizes the data handling operations in multi-hierarchy data structures, eliminates the need to wait for full aggregation to be complete, and provides for build-up of aggregated data required for full aggregation.
It is understood that the System and Method of the illustrative embodiments described herein above may be modified in a variety of ways which will become readily apparent to those skilled in the art of having the benefit of the novel teachings disclosed herein. All such modifications and variations of the illustrative embodiments thereof shall be deemed to be within the scope and spirit of the present invention as defined by the Claims to Invention appended hereto.
This is a Continuation of U.S. application Ser. No. 11/888,904 filed Aug. 2, 2007 now abandoned; which is a Continuation of U.S. application Ser. No. 10/839,782 filed May 5, 2004 now abandoned; which is a Continuation of U.S. application Ser. No. 10/314,884 filed Dec. 9, 2002, now U.S. Pat. No. 7,315,849; which is a Continuation of U.S. application Ser. No. 09/796,098 filed Feb. 28, 2001, now abandoned; which is a Continuation-in-part of: U.S. application Ser. No. 09/514,611 filed Feb. 28, 2000, now U.S. Pat. No. 6,434,544, and U.S. application Ser. No. 09/634,748 filed Aug. 9, 2000, now U.S. Pat. No. 6,385,604; each said Application being commonly owned by HyperRoll, Limited, and incorporated herein by reference in its entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4590465 | Fuchs | May 1986 | A |
| 4598400 | Hillis | Jul 1986 | A |
| 4641351 | Preston, Jr. | Feb 1987 | A |
| 4685144 | McCubbrey et al. | Aug 1987 | A |
| 4814980 | Peterson et al. | Mar 1989 | A |
| 4868733 | Fujisawa et al. | Sep 1989 | A |
| 4985834 | Cline et al. | Jan 1991 | A |
| 4985856 | Kaufman et al. | Jan 1991 | A |
| 4987554 | Kaufman | Jan 1991 | A |
| 4989141 | Lyons et al. | Jan 1991 | A |
| 5095427 | Tanaka et al. | Mar 1992 | A |
| 5101475 | Kaufman et al. | Mar 1992 | A |
| 5189608 | Lyons et al. | Feb 1993 | A |
| 5197005 | Shwartz et al. | Mar 1993 | A |
| 5202985 | Goyal | Apr 1993 | A |
| 5222216 | Parish et al. | Jun 1993 | A |
| 5222237 | Hillis | Jun 1993 | A |
| 5257365 | Powers et al. | Oct 1993 | A |
| 5278966 | Parks et al. | Jan 1994 | A |
| 5280474 | Nickolls et al. | Jan 1994 | A |
| 5293615 | Amada | Mar 1994 | A |
| 5297265 | Frank et al. | Mar 1994 | A |
| 5297280 | Potts et al. | Mar 1994 | A |
| 5299321 | Lizuka | Mar 1994 | A |
| 5307484 | Baker et al. | Apr 1994 | A |
| 5359724 | Earle | Oct 1994 | A |
| 5361385 | Bakalash | Nov 1994 | A |
| 5379419 | Heffernan et al. | Jan 1995 | A |
| 5381518 | Drebin et al. | Jan 1995 | A |
| 5386556 | Hedin et al. | Jan 1995 | A |
| 5404506 | Fujisawa et al. | Apr 1995 | A |
| 5410693 | Yu et al. | Apr 1995 | A |
| 5519859 | Grace | May 1996 | A |
| 5553226 | Kiuchi et al. | Sep 1996 | A |
| 5555408 | Fujisawa et al. | Sep 1996 | A |
| 5696916 | Yamazaki et al. | Dec 1997 | A |
| 5706495 | Chadha et al. | Jan 1998 | A |
| 5706503 | Poppen et al. | Jan 1998 | A |
| 5721910 | Unger et al. | Feb 1998 | A |
| 5742806 | Reiner et al. | Apr 1998 | A |
| 5745764 | Leach et al. | Apr 1998 | A |
| 5751928 | Bakalash | May 1998 | A |
| 5761652 | Wu et al. | Jun 1998 | A |
| 5765028 | Gladden | Jun 1998 | A |
| 5767854 | Anwar | Jun 1998 | A |
| 5781896 | Dalal | Jul 1998 | A |
| 5794228 | French et al. | Aug 1998 | A |
| 5794229 | French et al. | Aug 1998 | A |
| 5794246 | Sankaran et al. | Aug 1998 | A |
| 5799300 | Agrawal et al. | Aug 1998 | A |
| 5805885 | Leach et al. | Sep 1998 | A |
| 5822751 | Gray et al. | Oct 1998 | A |
| 5832475 | Agrawal et al. | Nov 1998 | A |
| 5848424 | Scheinkman et al. | Dec 1998 | A |
| 5850547 | Waddington et al. | Dec 1998 | A |
| 5852819 | Beller | Dec 1998 | A |
| 5852821 | Chen et al. | Dec 1998 | A |
| 5857184 | Lynch | Jan 1999 | A |
| 5864857 | Ohata et al. | Jan 1999 | A |
| 5867501 | Horst et al. | Feb 1999 | A |
| 5884299 | Ramesh et al. | Mar 1999 | A |
| 5890151 | Agrawal et al. | Mar 1999 | A |
| 5890154 | Hsiao et al. | Mar 1999 | A |
| 5901287 | Bull et al. | May 1999 | A |
| 5905985 | Malloy et al. | May 1999 | A |
| 5915257 | Gartung et al. | Jun 1999 | A |
| 5918225 | White et al. | Jun 1999 | A |
| 5918232 | Pouschine et al. | Jun 1999 | A |
| 5926818 | Malloy | Jul 1999 | A |
| 5926820 | Agrawal et al. | Jul 1999 | A |
| 5937410 | Shen | Aug 1999 | A |
| 5940818 | Malloy et al. | Aug 1999 | A |
| 5940822 | Haderle et al. | Aug 1999 | A |
| 5943668 | Malloy et al. | Aug 1999 | A |
| 5943677 | Hicks | Aug 1999 | A |
| 5946692 | Faloutsos et al. | Aug 1999 | A |
| 5946711 | Donnelly | Aug 1999 | A |
| 5963212 | Bakalash | Oct 1999 | A |
| 5963936 | Cochrane et al. | Oct 1999 | A |
| 5974416 | Anand et al. | Oct 1999 | A |
| 5978788 | Castelli et al. | Nov 1999 | A |
| 5978796 | Malloy et al. | Nov 1999 | A |
| 5987467 | Ross et al. | Nov 1999 | A |
| 5990892 | Urbain | Nov 1999 | A |
| 5991754 | Raitto et al. | Nov 1999 | A |
| 5999192 | Selfridge et al. | Dec 1999 | A |
| 6003024 | Bair et al. | Dec 1999 | A |
| 6003029 | Agrawal et al. | Dec 1999 | A |
| 6003036 | Martin | Dec 1999 | A |
| 6006216 | Griffin et al. | Dec 1999 | A |
| 6009432 | Tarin | Dec 1999 | A |
| 6014670 | Zamanian et al. | Jan 2000 | A |
| 6023695 | Osborn et al. | Feb 2000 | A |
| 6034697 | Becker | Mar 2000 | A |
| 6041103 | La Porta et al. | Mar 2000 | A |
| 6047323 | Krause | Apr 2000 | A |
| 6064999 | Dalal | May 2000 | A |
| 6073140 | Morgan et al. | Jun 2000 | A |
| 6078918 | Allen et al. | Jun 2000 | A |
| 6078924 | Ainsbury et al. | Jun 2000 | A |
| 6078994 | Carey | Jun 2000 | A |
| 6094651 | Agrawal et al. | Jul 2000 | A |
| 6108647 | Poosala et al. | Aug 2000 | A |
| 6115705 | Larson | Sep 2000 | A |
| 6115714 | Gallagher et al. | Sep 2000 | A |
| 6122628 | Castelli et al. | Sep 2000 | A |
| 6122636 | Malloy et al. | Sep 2000 | A |
| 6125624 | Prociw | Oct 2000 | A |
| 6134541 | Castelli et al. | Oct 2000 | A |
| 6141655 | Johnson et al. | Oct 2000 | A |
| 6151584 | Papierniak et al. | Nov 2000 | A |
| 6151601 | Papierniak et al. | Nov 2000 | A |
| 6154766 | Yost et al. | Nov 2000 | A |
| 6161103 | Rauer et al. | Dec 2000 | A |
| 6163774 | Lore et al. | Dec 2000 | A |
| 6167396 | Lokken | Dec 2000 | A |
| 6173310 | Yost et al. | Jan 2001 | B1 |
| 6182060 | Hedgcock et al. | Jan 2001 | B1 |
| 6182061 | Matsuzawa et al. | Jan 2001 | B1 |
| 6182062 | Fujisawa et al. | Jan 2001 | B1 |
| 6189004 | Rasen et al. | Feb 2001 | B1 |
| 6199063 | Colby et al. | Mar 2001 | B1 |
| 6205447 | Malloy | Mar 2001 | B1 |
| 6208975 | Bull et al. | Mar 2001 | B1 |
| 6209036 | Aldred et al. | Mar 2001 | B1 |
| 6212515 | Rogers | Apr 2001 | B1 |
| 6212524 | Weissman et al. | Apr 2001 | B1 |
| 6212617 | Hardwick | Apr 2001 | B1 |
| 6219654 | Ruffin | Apr 2001 | B1 |
| 6223573 | Grewal et al. | May 2001 | B1 |
| 6226647 | Venkatasu-bramanian et al. | May 2001 | B1 |
| 6249791 | Osborn et al. | Jun 2001 | B1 |
| 6256676 | Taylor et al. | Jul 2001 | B1 |
| 6260050 | Yost et al. | Jul 2001 | B1 |
| 6269393 | Yost et al. | Jul 2001 | B1 |
| 6275818 | Subramanian et al. | Aug 2001 | B1 |
| 6282544 | Tse et al. | Aug 2001 | B1 |
| 6285994 | Bui et al. | Sep 2001 | B1 |
| 6289334 | Reiner et al. | Sep 2001 | B1 |
| 6289352 | Proctor | Sep 2001 | B1 |
| 6301579 | Becker | Oct 2001 | B1 |
| 6317750 | Tortolani et al. | Nov 2001 | B1 |
| 6321206 | Honarvar | Nov 2001 | B1 |
| 6321241 | Gartung et al. | Nov 2001 | B1 |
| 6324533 | Agrawal et al. | Nov 2001 | B1 |
| 6324623 | Carey | Nov 2001 | B1 |
| 6330564 | Hellerstein et al. | Dec 2001 | B1 |
| 6332130 | Notani et al. | Dec 2001 | B1 |
| 6339775 | Zamanian et al. | Jan 2002 | B1 |
| 6356900 | Egilsson et al. | Mar 2002 | B1 |
| 6363353 | Chen | Mar 2002 | B1 |
| 6363393 | Ribitzky | Mar 2002 | B1 |
| 6366905 | Netz | Apr 2002 | B1 |
| 6366922 | Althoff | Apr 2002 | B1 |
| 6374234 | Netz | Apr 2002 | B1 |
| 6374263 | Bunger et al. | Apr 2002 | B1 |
| 6377934 | Chen et al. | Apr 2002 | B1 |
| 6381605 | Kothuri et al. | Apr 2002 | B1 |
| 6385301 | Nolting et al. | May 2002 | B1 |
| 6385604 | Bakalash et al. | May 2002 | B1 |
| 6397195 | Pinard et al. | May 2002 | B1 |
| 6401117 | Narad et al. | Jun 2002 | B1 |
| 6405173 | Honarvar et al. | Jun 2002 | B1 |
| 6405207 | Petculescu et al. | Jun 2002 | B1 |
| 6405208 | Raghavan et al. | Jun 2002 | B1 |
| 6408292 | Bakalash et al. | Jun 2002 | B1 |
| 6411313 | Conlon et al. | Jun 2002 | B1 |
| 6411681 | Nolting et al. | Jun 2002 | B1 |
| 6411961 | Chen et al. | Jun 2002 | B1 |
| 6418427 | Egilsson et al. | Jul 2002 | B1 |
| 6418450 | Daudenarde | Jul 2002 | B2 |
| 6421730 | Narad et al. | Jul 2002 | B1 |
| 6424979 | Livingston et al. | Jul 2002 | B1 |
| 6430545 | Honarvar et al. | Aug 2002 | B1 |
| 6430547 | Busche et al. | Aug 2002 | B1 |
| 6434544 | Bakalash et al. | Aug 2002 | B1 |
| 6434557 | Egilsson et al. | Aug 2002 | B1 |
| 6438537 | Netz et al. | Aug 2002 | B1 |
| 6441834 | Agassi et al. | Aug 2002 | B1 |
| 6442269 | Ehrlich et al. | Aug 2002 | B1 |
| 6442560 | Berger et al. | Aug 2002 | B1 |
| 6446059 | Berger et al. | Sep 2002 | B1 |
| 6446061 | Doerre et al. | Sep 2002 | B1 |
| 6453322 | DeKimpe et al. | Sep 2002 | B1 |
| 6456999 | Netz | Sep 2002 | B1 |
| 6460026 | Pasumansky | Oct 2002 | B1 |
| 6460031 | Wilson et al. | Oct 2002 | B1 |
| 6470344 | Kothuri et al. | Oct 2002 | B1 |
| 6473750 | Petculescu et al. | Oct 2002 | B1 |
| 6473764 | Petculescu et al. | Oct 2002 | B1 |
| 6477536 | Pasumansky et al. | Nov 2002 | B1 |
| 6480842 | Agassi et al. | Nov 2002 | B1 |
| 6480848 | DeKimpe et al. | Nov 2002 | B1 |
| 6480850 | Veldhuisen | Nov 2002 | B1 |
| 6484179 | Roccaforte | Nov 2002 | B1 |
| 6487547 | Ellison et al. | Nov 2002 | B1 |
| 6493718 | Petculescu et al. | Dec 2002 | B1 |
| 6493723 | Busche | Dec 2002 | B1 |
| 6493728 | Berger | Dec 2002 | B1 |
| 6510457 | Ayukawa et al. | Jan 2003 | B1 |
| 6513019 | Lewis | Jan 2003 | B2 |
| 6532458 | Chaudhuri et al. | Mar 2003 | B1 |
| 6535866 | Iwadate | Mar 2003 | B1 |
| 6535868 | Galeazzi et al. | Mar 2003 | B1 |
| 6535872 | Castelli et al. | Mar 2003 | B1 |
| 6542886 | Chaudhuri et al. | Apr 2003 | B1 |
| 6542895 | DeKimpe et al. | Apr 2003 | B1 |
| 6545589 | Fuller et al. | Apr 2003 | B1 |
| 6546395 | DeKimpe et al. | Apr 2003 | B1 |
| 6546545 | Honarvar et al. | Apr 2003 | B1 |
| 6549907 | Fayyad et al. | Apr 2003 | B1 |
| 6557008 | Temple et al. | Apr 2003 | B1 |
| 6560594 | Cochrane et al. | May 2003 | B2 |
| 6567796 | Yost et al. | May 2003 | B1 |
| 6567814 | Bankier et al. | May 2003 | B1 |
| 6581054 | Bogrett | Jun 2003 | B1 |
| 6581068 | Bensoussan et al. | Jun 2003 | B1 |
| 6587547 | Zirngibl et al. | Jul 2003 | B1 |
| 6587857 | Carothers et al. | Jul 2003 | B1 |
| 6594672 | Lampson et al. | Jul 2003 | B1 |
| 6601034 | Honarvar et al. | Jul 2003 | B1 |
| 6601062 | Deshpande | Jul 2003 | B1 |
| 6604135 | Rogers et al. | Aug 2003 | B1 |
| 6606638 | Tarin | Aug 2003 | B1 |
| 6609120 | Honarvar et al. | Aug 2003 | B1 |
| 6615096 | Durrant et al. | Sep 2003 | B1 |
| 6628312 | Rao et al. | Sep 2003 | B1 |
| 6629094 | Colby | Sep 2003 | B1 |
| 6633875 | Brady | Oct 2003 | B2 |
| 6636870 | Roccaforte | Oct 2003 | B2 |
| 6643608 | Hershey et al. | Nov 2003 | B1 |
| 6643661 | Polizzi et al. | Nov 2003 | B2 |
| 6662174 | Shah et al. | Dec 2003 | B2 |
| 6665682 | DeKimpe et al. | Dec 2003 | B1 |
| 6671715 | Langseth et al. | Dec 2003 | B1 |
| 6677963 | Mani et al. | Jan 2004 | B1 |
| 6678674 | Saeki | Jan 2004 | B1 |
| 6691118 | Gongwer et al. | Feb 2004 | B1 |
| 6691140 | Bogrett | Feb 2004 | B1 |
| 6694316 | Langseth et al. | Feb 2004 | B1 |
| 6707454 | Barg et al. | Mar 2004 | B1 |
| 6708155 | Honarvar et al. | Mar 2004 | B1 |
| 6732115 | Shah et al. | May 2004 | B2 |
| 6738975 | Vee et al. | May 2004 | B1 |
| 6748394 | Shah et al. | Jun 2004 | B2 |
| 6763357 | Deshpande et al. | Jul 2004 | B1 |
| 6766325 | Pasumansky et al. | Jul 2004 | B1 |
| 6775674 | Agassi et al. | Aug 2004 | B1 |
| 6778996 | Roccaforte | Aug 2004 | B2 |
| 6801908 | Fuloria et al. | Oct 2004 | B1 |
| 6816854 | Reiner et al. | Nov 2004 | B2 |
| 6826593 | Acharya et al. | Nov 2004 | B1 |
| 6832263 | Polizzi et al. | Dec 2004 | B2 |
| 6836894 | Hellerstein et al. | Dec 2004 | B1 |
| 6842758 | Bogrett | Jan 2005 | B1 |
| 6867788 | Takeda | Mar 2005 | B1 |
| 6898603 | Petculescu et al. | May 2005 | B1 |
| 6934687 | Papierniak et al. | Aug 2005 | B1 |
| 6947934 | Chen et al. | Sep 2005 | B1 |
| 7096219 | Karch | Aug 2006 | B1 |
| 7315849 | Bakalash et al. | Jan 2008 | B2 |
| 7333982 | Bakalash et al. | Feb 2008 | B2 |
| 7392248 | Bakalash et al. | Jun 2008 | B2 |
| 7529730 | Potter et al. | May 2009 | B2 |
| 7778899 | Scumniotales et al. | Aug 2010 | B2 |
| 7853508 | Scumniotales et al. | Dec 2010 | B2 |
| 8041670 | Bakalash et al. | Oct 2011 | B2 |
| 20010013030 | Colby et al. | Aug 2001 | A1 |
| 20020016924 | Shah et al. | Feb 2002 | A1 |
| 20020023122 | Polizzi et al. | Feb 2002 | A1 |
| 20020029207 | Bakalash et al. | Mar 2002 | A1 |
| 20020038229 | Shah et al. | Mar 2002 | A1 |
| 20020038297 | Shah et al. | Mar 2002 | A1 |
| 20020077997 | Colby et al. | Jun 2002 | A1 |
| 20020091707 | Keller | Jul 2002 | A1 |
| 20020099692 | Shah et al. | Jul 2002 | A1 |
| 20020129003 | Bakalash et al. | Sep 2002 | A1 |
| 20020129032 | Bakalash et al. | Sep 2002 | A1 |
| 20020143783 | Bakalash et al. | Oct 2002 | A1 |
| 20020184187 | Bakalash et al. | Dec 2002 | A1 |
| 20020194167 | Bakalash et al. | Dec 2002 | A1 |
| 20030018642 | Bakalash et al. | Jan 2003 | A1 |
| 20030055832 | Roccaforte | Mar 2003 | A1 |
| 20030200221 | Bakalash et al. | Oct 2003 | A1 |
| 20030208503 | Roccaforte | Nov 2003 | A1 |
| 20030217079 | Bakalash et al. | Nov 2003 | A1 |
| 20030225736 | Bakalash et al. | Dec 2003 | A1 |
| 20030225752 | Bakalash et al. | Dec 2003 | A1 |
| 20030229652 | Bakalash et al. | Dec 2003 | A1 |
| 20040073566 | Trivedi | Apr 2004 | A1 |
| 20040236655 | Scumniotales et al. | Nov 2004 | A1 |
| 20040243607 | Tummalapalli | Dec 2004 | A1 |
| 20040247105 | Mullis et al. | Dec 2004 | A1 |
| 20050038799 | Jordan et al. | Feb 2005 | A1 |
| 20050055329 | Bakalash et al. | Mar 2005 | A1 |
| 20050060325 | Bakalash et al. | Mar 2005 | A1 |
| 20050060326 | Bakalash et al. | Mar 2005 | A1 |
| 20050065940 | Bakalash et al. | Mar 2005 | A1 |
| 20050076067 | Bakalash et al. | Apr 2005 | A1 |
| 20050091237 | Bakalash et al. | Apr 2005 | A1 |
| 20050114243 | Scumniotales et al. | May 2005 | A1 |
| 20050149491 | Bakalash et al. | Jul 2005 | A1 |
| 20070192295 | Bakalash et al. | Aug 2007 | A1 |
| 20070233644 | Bakalash et al. | Oct 2007 | A1 |
| 20080016043 | Bakalash et al. | Jan 2008 | A1 |
| 20080016057 | Bakalash et al. | Jan 2008 | A1 |
| 20080021864 | Bakalash et al. | Jan 2008 | A1 |
| 20080021893 | Bakalash et al. | Jan 2008 | A1 |
| 20080021915 | Bakalash et al. | Jan 2008 | A1 |
| 20080059415 | Bakalash et al. | Mar 2008 | A1 |
| 20080211817 | Bakalash et al. | Sep 2008 | A1 |
| 20090076983 | Scumniotales et al. | Mar 2009 | A1 |
| 20090077107 | Scumniotales et al. | Mar 2009 | A1 |
| 20090271379 | Bakalash et al. | Oct 2009 | A1 |
| 20090271384 | Bakalash et al. | Oct 2009 | A1 |
| 20090276410 | Bakalash et al. | Nov 2009 | A1 |
| 20100042645 | Bakalash et al. | Feb 2010 | A1 |
| 20100063958 | Bakalash et al. | Mar 2010 | A1 |
| 20100100558 | Bakalash et al. | Apr 2010 | A1 |
| 20100185581 | Bakalash et al. | Jul 2010 | A1 |
| Number | Date | Country |
|---|---|---|
| 0 314 279 | May 1989 | EP |
| 0 657 052 | Jun 1995 | EP |
| 0 743 609 | Nov 1996 | EP |
| 0336 584 | Feb 1997 | EP |
| 0 869 444 | Oct 1998 | EP |
| 1 266 308 | Dec 2002 | EP |
| 9-265479 | Oct 1997 | JP |
| 2001-565050 | Feb 2001 | JP |
| WO 9119269 | Dec 1991 | WO |
| WO 9404991 | Mar 1994 | WO |
| WO 9508794 | Mar 1995 | WO |
| WO 9822908 | May 1998 | WO |
| WO 9822908 | May 1998 | WO |
| WO 9840829 | Sep 1998 | WO |
| WO 9849636 | Nov 1998 | WO |
| WO 9909492 | Feb 1999 | WO |
| WO 01011497 | Feb 2001 | WO |
| WO 0167303 | Sep 2001 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 20090271379 A1 | Oct 2009 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11888904 | Aug 2007 | US |
| Child | 12455664 | US | |
| Parent | 10839782 | May 2004 | US |
| Child | 11888904 | US | |
| Parent | 10314884 | Dec 2002 | US |
| Child | 10839782 | US | |
| Parent | 09796098 | Feb 2001 | US |
| Child | 10314884 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 09514611 | Feb 2000 | US |
| Child | 09796098 | US | |
| Parent | 09634748 | Aug 2000 | US |
| Child | 09514611 | US |
