An Overview of Data Warehousing and OLAP Technology
Usman Ahmad Urfi
Mphil CS Lahore Leads University 2nd Semester, Lahore
Semester, Lahore [email protected]
Submitted to:- Most Respectful Madam Sadaf Ali
Data warehousing and on-line analytical processing (OLAP) are essential elements of decision support, which has increasingly become a focus of the database industry. Many commercial products and services are now available, and all of the principal database management system vendors now have offerings in these areas. Decision support places some rather different requirements on database technology compared to traditional on-line transaction processing applications. This paper provides an overview of data warehousing and OLAP technologies, with an emphasis on their new requirements. We describe back end tools for extracting, cleaning and loading data into a data warehouse; multidimensional data models typical of OLAP; front end client tools for querying and data analysis; server extensions for efficient query processing; and tools for metadata management and for managing the warehouse. In addition to surveying the state of the art, this paper also identifies some promising research issues, some of which are related to problems that the database research community has worked on for years, but others are only just beginning to be addressed. This overview is based on a tutorial that the authors presented at the VLDB Conference, 1996.
Data warehousing is a collection of decision support technologies, aimed at enabling the knowledge worker (executive, manager, analyst) to make better and faster decisions. The past three years have seen explosive growth, both in the number of products and services offered, and in the adoption of these technologies by industry. According to the META Group, the data warehousing market, including hardware, database software, and tools, is projected to grow from $2 billion in 1995 to $8 billion in 1998. Data warehousing technologies have been successfully deployed in many industries: manufacturing (for order shipment and customer support), retail (for user profiling and inventory management), financial services (for claims analysis, risk analysis, credit card analysis, and fraud detection), transportation (for fleet management), telecommunications (for call analysis and fraud detection), utilities (for power usage analysis), and healthcare (for outcomes analysis). This paper presents a roadmap of data warehousing technologies, focusing on the special requirements that data warehouses place on database management systems (DBMSs).
A data warehouse is a “subject-oriented, integrated, time- varying, non-volatile collection of data that is used primarily in organizational decision making.”1 Typically, the data warehouse is maintained separately from the organization’s operational databases. There are many reasons for doing this. The data warehouse supports on-line analytical processing (OLAP), the functional and performance requirements of which are quite different from those of the on-line transaction processing (OLTP) applications traditionally supported by the operational databases.
OLTP applications typically automate clerical data processing tasks such as order entry and banking transactions that are the bread-and-butter day-to-day operations of an organization. These tasks are structured and repetitive, and consist of short, atomic, isolated transactions. The transactions require detailed, up-to-date data, and read or update a few (tens of) records accessed typically on their primary keys. Operational databases tend to be hundreds of megabytes to gigabytes in size. Consistency and recoverability of the database are critical, and maximizing transaction throughput is the key performance metric. Consequently, the database is designed to reflect the operational semantics of known applications, and, in particular, to minimize concurrency conflicts.
Data warehouses, in contrast, are targeted for decision support. Historical, summarized and consolidated data is more important than detailed, individual records. Since data warehouses contain consolidated data, perhaps from several operational databases, over potentially long periods of time, they tend to be orders of magnitude larger than operational databases; enterprise data warehouses are projected to be hundreds of gigabytes to terabytes in size. The workloads are query intensive with mostly ad hoc, complex queries that can access millions of records and perform a lot of scans, joins, and aggregates. Query throughput and response times are more important than transaction throughput.
To facilitate complex analyses and visualization, the data in a warehouse is typically modeled multidimensionally. For example, in a sales data warehouse, time of sale, sales district, salesperson, and product might be some of the dimensions of interest. Often, these dimensions are hierarchical; time of sale may be organized as a day-month-quarter-year hierarchy, product as a product-category-industry hierarchy. Typical
OLAP operations include rollup (increasing the level of aggregation) and drill-down (decreasing the level of aggregation or increasing detail) along one or more dimension hierarchies, slice_and_dice (selection and projection), and pivot (re-orienting the multidimensional view of data).
Given that operational databases are finely tuned to support known OLTP workloads, trying to execute complex OLAP queries against the operational databases would result in unacceptable performance. Furthermore, decision support requires data that might be missing from the operational databases; for instance, understanding trends or making predictions requires historical data, whereas operational databases store only current data. Decision support usually requires consolidating data from many heterogeneous sources: these might include external sources such as stock market feeds, in addition to several operational databases. The different sources might contain data of varying quality, or use inconsistent representations, codes and formats, which have to be reconciled. Finally, supporting the multidimensional data models and operations typical of OLAP requires special data organization, access methods, and implementation methods, not generally provided by commercial DBMSs targeted for OLTP. It is for all these reasons that data warehouses are implemented separately from operational databases.
Data warehouses might be implemented on standard or extended relational DBMSs, called Relational OLAP (ROLAP) servers. These servers assume that data is stored in relational databases, and they support extensions to SQL and special access and implementation methods to efficiently
In Section 2, we describe a typical data warehousing architecture, and the process of designing and operating a data warehouse. In Sections 3-7, we review relevant technologies for loading and refreshing data in a data warehouse, warehouse servers, front end tools, and warehouse management tools. In each case, we point out what is different from traditional database technology, and we mention representative products. In this paper, we do not intend to provide comprehensive descriptions of all products in every category. We encourage the interested reader to look at recent issues of trade magazines such as Databased Advisor, Database Programming and Design, Datamation, and DBMS Magazine, and vendors’ Web sites for more details of commercial products, white papers, and case studies. The OLAP Council2 is a good source of information on standardization efforts across the industry, and a paper by Codd, et al.3 defines twelve rules for OLAP products. Finally, a good source of references on data warehousing and OLAP is the Data Warehousing Information Center4.
Research in data warehousing is fairly recent, and has focused primarily on query processing and view maintenance issues. There still are many open research problems. We conclude in Section 8 with a brief mention of these issues.
2. Architecture and End-to-End Process
Figure 1 shows a typical data warehousing architecture.
Monitoring & Admnistration
implement the multidimensional data model and operations. In contrast, multidimensional OLAP (MOLAP) servers are servers that directly store multidimensional data in special data structures (e.g., arrays) and implement the OLAP operations over these special data structures.
Extract Transform Load Refresh
There is more to building and maintaining a data warehouse than selecting an OLAP server and defining a schema and some complex queries for the warehouse. Different architectural alternatives exist. Many organizations want to
Figure 1. Data Warehousing Architecture
implement an integrated enterprise warehouse that collects information about all subjects (e.g., customers, products, sales, assets, personnel) spanning the whole organization. However, building an enterprise warehouse is a long and complex process, requiring extensive business modeling, and may take many years to succeed. Some organizations are settling for data marts instead, which are departmental subsets focused on selected subjects (e.g., a marketing data mart may include customer, product, and sales information). These data marts enable faster roll out, since they do not require enterprise-wide consensus, but they may lead to complex integration problems in the long run, if a complete business model is not developed.
It includes tools for extracting data from multiple operational databases and external sources; for cleaning, transforming and integrating this data; for loading data into the data warehouse; and for periodically refreshing the warehouse to reflect updates at the sources and to purge data from the warehouse, perhaps onto slower archival storage. In addition to the main warehouse, there may be several departmental data marts. Data in the warehouse and data marts is stored and managed by one or more warehouse servers, which present multidimensional views of data to a variety of front end tools: query tools, report writers, analysis tools, and data mining tools. Finally, there is a repository for storing and
managing metadata, and tools for monitoring and administering the warehousing system.
The warehouse may be distributed for load balancing, scalability, and higher availability. In such a distributed architecture, the metadata repository is usually replicated with each fragment of the warehouse, and the entire warehouse is administered centrally. An alternative architecture, implemented for expediency when it may be too expensive to construct a single logically integrated enterprise warehouse, is a federation of warehouses or data marts, each with its own repository and decentralized administration.
Designing and rolling out a data warehouse is a complex process, consisting of the following activities5.
• Define the architecture, do capacity planning, and select the storage servers, database and OLAP servers, and tools.
• Integrate the servers, storage, and client tools.
• Design the warehouse schema and views.
• Define the physical warehouse organization, data placement, partitioning, and access methods.
• Connect the sources using gateways, ODBC drivers, or other wrappers.
• Design and implement scripts for data extraction, cleaning, transformation, load, and refresh.
• Populate the repository with the schema and view definitions, scripts, and other metadata.
• Design and implement end-user applications.
• Roll out the warehouse and applications.
3. Back End Tools and Utilities
Data warehousing systems use a variety of data extraction and cleaning tools, and load and refresh utilities for populating warehouses. Data extraction from “foreign” sources is usually implemented via gateways and standard interfaces (such as Information Builders EDA/SQL, ODBC, Oracle Open Connect, Sybase Enterprise Connect, Informix Enterprise Gateway).
Since a data warehouse is used for decision making, it is important that the data in the warehouse be correct. However, since large volumes of data from multiple sources are involved, there is a high probability of errors and anomalies in the data.. Therefore, tools that help to detect data anomalies and correct them can have a high payoff. Some examples where data cleaning becomes necessary are: inconsistent field lengths, inconsistent descriptions, inconsistent value assignments, missing entries and violation of integrity constraints. Not surprisingly, optional fields in data entry forms are significant sources of inconsistent data.
There are three related, but somewhat different, classes of data cleaning tools. Data migration tools allow simple transformation rules to be specified; e.g., “replace the string gender by sex”. Warehouse Manager from Prism is an example of a popular tool of this kind. Data scrubbing tools use domain-specific knowledge (e.g., postal addresses) to do the scrubbing of data. They often exploit parsing and fuzzy matching techniques to accomplish cleaning from multiple sources. Some tools make it possible to specify the “relative cleanliness” of sources. Tools such as Integrity and Trillum fall in this category. Data auditing tools make it possible to discover rules and relationships (or to signal violation of stated rules) by scanning data. Thus, such tools may be considered variants of data mining tools. For example, such a tool may discover a suspicious pattern (based on statistical analysis) that a certain car dealer has never received any complaints.
After extracting, cleaning and transforming, data must be loaded into the warehouse. Additional preprocessing may still be required: checking integrity constraints; sorting; summarization, aggregation and other computation to build the derived tables stored in the warehouse; building indices and other access paths; and partitioning to multiple target storage areas. Typically, batch load utilities are used for this purpose. In addition to populating the warehouse, a load utility must allow the system administrator to monitor status, to cancel, suspend and resume a load, and to restart after failure with no loss of data integrity.
The load utilities for data warehouses have to deal with much larger data volumes than for operational databases. There is only a small time window (usually at night) when the warehouse can be taken offline to refresh it. Sequential loads can take a very long time, e.g., loading a terabyte of data can take weeks and months! Hence, pipelined and partitioned parallelism are typically exploited 6. Doing a full load has the advantage that it can be treated as a long batch transaction that builds up a new database. While it is in progress, the current database can still support queries; when the load transaction commits, the current database is replaced with the new one. Using periodic checkpoints ensures that if a failure occurs during the load, the process can restart from the last checkpoint.
However, even using parallelism, a full load may still take too long. Most commercial utilities (e.g., RedBrick Table Management Utility) use incremental loading during refresh to reduce the volume of data that has to be incorporated into the warehouse. Only the updated tuples are inserted. However, the load process now is harder to manage. The incremental load conflicts with ongoing queries, so it is treated as a sequence of shorter transactions (which commit periodically, e.g., after every 1000 records or every few seconds), but now this sequence of transactions has to be
coordinated to ensure consistency of derived data and indices with the base data.
Refreshing a warehouse consists in propagating updates on source data to correspondingly update the base data and derived data stored in the warehouse. There are two sets of issues to consider: when to refresh, and how to refresh. Usually, the warehouse is refreshed periodically (e.g., daily or weekly). Only if some OLAP queries need current data (e.g., up to the minute stock quotes), is it necessary to propagate every update. The refresh policy is set by the warehouse administrator, depending on user needs and traffic, and may be different for different sources.
Refresh techniques may also depend on the characteristics of the source and the capabilities of the database servers. Extracting an entire source file or database is usually too expensive, but may be the only choice for legacy data sources. Most contemporary database systems provide replication servers that support incremental techniques for propagating updates from a primary database to one or more replicas. Such replication servers can be used to incrementally refresh a warehouse when the sources change. There are two basic replication techniques: data shipping and transaction shipping.
In data shipping (e.g., used in the Oracle Replication Server, Praxis OmniReplicator), a table in the warehouse is treated as a remote snapshot of a table in the source database. After_row triggers are used to update a snapshot log table whenever the source table changes; and an automatic refresh schedule (or a manual refresh procedure) is then set up to propagate the updated data to the remote snapshot.
correct updates for incrementally updating derived data (materialized views) has been the subject of much research 7 8
9 10. For data warehousing, the most significant classes of
derived data are summary tables, single-table indices and join indices.
4. Conceptual Model and Front End Tools
A popular conceptual model that influences the front-end tools, database design, and the query engines for OLAP is the multidimensional view of data in the warehouse. In a multidimensional data model, there is a set of numeric measures that are the objects of analysis. Examples of such measures are sales, budget, revenue, inventory, ROI (return on investment). Each of the numeric measures depends on a set of dimensions, which provide the context for the measure. For example, the dimensions associated with a sale amount can be the city, product name, and the date when the sale was made. The dimensions together are assumed to uniquely determine the measure. Thus, the multidimensional data views a measure as a value in the multidimensional space of dimensions. Each dimension is described by a set of attributes. For example, the Product dimension may consist of four attributes: the category and the industry of the product, year of its introduction, and the average profit margin. For example, the soda Surge belongs to the category beverage and the food industry, was introduced in 1996, and may have an average profit margin of 80%. The attributes of a dimension may be related via a hierarchy of relationships. In the above example, the product name is related to its category and the industry attribute through such a hierarchical relationship.
In transaction shipping (e.g., used in the Sybase Replication Server and Microsoft SQL Server), the regular transaction log is used, instead of triggers and a special snapshot log table. At the source site, the transaction log is sniffed to detect updates on replicated tables, and those log records are transferred to a replication server, which packages up the corresponding transactions to update the replicas. Transaction shipping has the advantage that it does not require triggers, which can increase the workload on the operational source
Milk Cream Toothpaste
1 2 3 4 5 6 7
Dimensions: Product, City, Date
Hierarchical summarization paths
Industry Country Year Category State Quarter Product City Month Week
databases. However, it cannot always be used easily across DBMSs from different vendors, because there are no standard APIs for accessing the transaction log.
Such replication servers have been used for refreshing data warehouses. However, the refresh cycles have to be properly chosen so that the volume of data does not overwhelm the incremental load utility.
In addition to propagating changes to the base data in the warehouse, the derived data also has to be updated correspondingly. The problem of constructing logically
Figure 2. Multidimensional data
Another distinctive feature of the conceptual model for OLAP is its stress on aggregation of measures by one or more dimensions as one of the key operations; e.g., computing and ranking the total sales by each county (or by each year). Other popular operations include comparing two measures (e.g., sales and budget) aggregated by the same dimensions. Time is a dimension that is of particular significance to decision support (e.g., trend analysis). Often, it is desirable to have built-in knowledge of calendars and other aspects of the time dimension.
Front End Tools
The multidimensional data model grew out of the view of business data popularized by PC spreadsheet programs that were extensively used by business analysts. The spreadsheet is still the most compelling front-end application for OLAP. The challenge in supporting a query environment for OLAP can be crudely summarized as that of supporting spreadsheet operations efficiently over large multi-gigabyte databases. Indeed, the Essbase product of Arbor Corporation uses Microsoft Excel as the front-end tool for its multidimensional engine.
We shall briefly discuss some of the popular operations that are supported by the multidimensional spreadsheet applications. One such operation is pivoting. Consider the multidimensional schema of Figure 2 represented in a spreadsheet where each row corresponds to a sale . Let there be one column for each dimension and an extra column that represents the amount of sale. The simplest view of pivoting is that it selects two dimensions that are used to aggregate a measure, e.g., sales in the above example. The aggregated values are often displayed in a grid where each value in the (x,y) coordinate corresponds to the aggregated value of the measure when the first dimension has the value x and the second dimension has the value y. Thus, in our example, if the selected dimensions are city and year, then the x-axis may represent all values of city and the y-axis may represent the years. The point (x,y) will represent the aggregated sales for city x in the year y. Thus, what were values in the original spreadsheets have now become row and column headers in the pivoted spreadsheet.
Other operators related to pivoting are rollup or drill-down.
Rollup corresponds to taking the current data object and
data. These applications often use raw data access tools and optimize the access patterns depending on the back end database server. In addition, there are query environments (e.g., Microsoft Access) that help build ad hoc SQL queries by “pointing-and-clicking”. Finally, there are a variety of data mining tools that are often used as front end tools to data warehouses.
5. Database Design Methodology
The multidimensional data model described above is implemented directly by MOLAP servers. We will describe these briefly in the next section. However, when a relational ROLAP server is used, the multidimensional model and its operations have to be mapped into relations and SQL queries. In this section, we describe the design of relational database schemas that reflect the multidimensional views of data.
Entity Relationship diagrams and normalization techniques are popularly used for database design in OLTP environments. However, the database designs recommended by ER diagrams are inappropriate for decision support systems where efficiency in querying and in loading data (including incremental loads) are important.
Most data warehouses use a star schema to represent the multidimensional data model. The database consists of a single fact table and a single table for each dimension. Each tuple in the fact table consists of a pointer (foreign key – often uses a generated key for efficiency) to each of the dimensions that provide its multidimensional coordinates, and stores the numeric measures for those coordinates. Each dimension table consists of columns that correspond to attributes of the dimension. Figure 3 shows an example of a star schema.
doing a further group-by on one of the dimensions. Thus, it is possible to roll-up the sales data, perhaps already aggregated on city, additionally by product. The drill-down operation is the converse of rollup. Slice_and_dice corresponds to reducing the dimensionality of the data, i.e., taking a projection of the data on a subset of dimensions for selected values of the other dimensions. For example, we can slice_and_dice sales data for a specific product to create a table that consists of the dimensions city and the day of sale. The other popular operators include ranking (sorting), selections and defining computed attributes.
Order OrderNo OrderDate
CustomerNo CustomerName CustomerAddress City
Salesperson SalespersonID SalespesonName
OrderNo SalespersonID CustomerNo ProdNo DateKey CityName Quantity TotalPrice
ProdNo ProdName ProdDescr Category CategoryDescr UnitPrice QOH
DateKey Date Month Year
Although the multidimensional spreadsheet has attracted a lot
of interest since it empowers the end user to analyze business data, this has not replaced traditional analysis by means of a managed query environment. These environments use stored procedures and predefined complex queries to provide
Figure 3. A Star Schema.
packaged analysis tools. Such tools often make it possible for the end-user to query in terms of domain-specific business
Star schemas do not explicitly provide support for attribute
hierarchies. Snowflake schemas provide a refinement of star
schemas where the dimensional hierarchy is explicitly represented by normalizing the dimension tables, as shown in Figure 4. This leads to advantages in maintaining the dimension tables. However, the denormalized structure of the dimensional tables in star schemas may be more appropriate for browsing the dimensions.
Fact constellations are examples of more complex structures in which multiple fact tables share dimensional tables. For example, projected expense and the actual expense may form a fact constellation since they share many dimensions.
Data warehouses may contain large volumes of data. To answer queries efficiently, therefore, requires highly efficient access methods and query processing techniques. Several issues arise. First, data warehouses use redundant structures such as indices and materialized views. Choosing which indices to build and which views to materialize is an important physical design problem. The next challenge is to effectively use the existing indices and materialized views to answer queries. Optimization of complex queries is another important problem. Also, while for data-selective queries, efficient index scans may be very effective, data-intensive
Order OrderNo OrderDate
queries need the use of sequential scans. Thus, improving the efficiency of scans is important. Finally, parallelism needs to be exploited to reduce query response times. In this short paper, it is not possible to elaborate on each of these issues. Therefore, we will only briefly touch upon the highlights.
Salesperson SalespersonID SalespesonName City
Index Structures and their Usage
A number of query processing techniques that exploit indices are useful. For instance, the selectivities of multiple conditions can be exploited through index intersection. Other useful index operations are union of indexes. These index
Figure 4. A Snowflake Schema.
In addition to the fact and dimension tables, data warehouses store selected summary tables containing pre-aggregated data. In the simplest cases, the pre-aggregated data corresponds to aggregating the fact table on one or more selected dimensions. Such pre-aggregated summary data can be represented in the database in at least two ways. Let us consider the example of a summary table that has total sales by product by year in the context of the star schema of Figure
3. We can represent such a summary table by a separate fact table which shares the dimension Product and also a separate shrunken dimension table for time, which consists of only the attributes of the dimension that make sense for the summary table (i.e., year). Alternatively, we can represent the summary table by encoding the aggregated tuples in the same fact table and the same dimension tables without adding new tables. This may be accomplished by adding a new level field to each dimension and using nulls: We can encode a day, a month or a year in the Date dimension table as follows: (id0, 0, 22, 01, 1960) represents a record for Jan 22, 1960, (id1, 1, NULL, 01, 1960) represents the month Jan 1960 and (id2, 2, NULL, NULL, 1960) represents the year 1960. The second attribute represents the new attribute level: 0 for days, 1 for months, 2 for years. In the fact table, a record containing the foreign key id2 represents the aggregated sales for a Product in the year 1960. The latter method, while reducing the number of tables, is often a source of operational errors since the level field needs be carefully interpreted.
6. Warehouse Servers
operations can be used to significantly reduce and in many cases eliminate the need to access the base tables.
Warehouse servers can use bit map indices, which support efficient index operations (e.g., union, intersection). Consider a leaf page in an index structure corresponding to a domain value d. Such a leaf page traditionally contains a list of the record ids (RIDs) of records that contain the value d. However, bit map indices use an alternative representation of the above RID list as a bit vector that has one bit for each record, which is set when the domain value for that record is
d. In a sense, the bit map index is not a new index structure, but simply an alternative representation of the RID list. The popularity of the bit map index is due to the fact that the bit vector representation of the RID lists can speed up index intersection, union, join, and aggregation11. For example, if we have a query of the form column1 = d & column2 = d’, then we can identify the qualifying records by taking the AND of the two bit vectors. While such representations can be very useful for low cardinality domains (e.g., gender), they can also be effective for higher cardinality domains through compression of bitmaps (e.g., run length encoding). Bitmap indices were originally used in Model 204, but many products support them today (e.g., Sybase IQ). An interesting question is to decide on which attributes to index. In general, this is really a question that must be answered by the physical database design process.
In addition to indices on single tables, the specialized nature of star schemas makes join indices especially attractive for decision support. While traditionally indices map the value in a column to a list of rows with that value, a join index
maintains the relationships between a foreign key with its matching primary keys. In the context of a star schema, a join index can relate the values of one or more attributes of a dimension table to matching rows in the fact table. For example, consider the schema of Figure 3. There can be a join index on City that maintains, for each city, a list of RIDs of the tuples in the fact table that correspond to sales in that city. Thus a join index essentially precomputes a binary join. Multikey join indices can represent precomputed n-way joins. For example, over the Sales database it is possible to construct a multidimensional join index from (Cityname, Productname) to the fact table. Thus, the index entry for (Seattle, jacket) points to RIDs of those tuples in the Sales table that have the above combination. Using such a multidimensional join index can sometimes provide savings over taking the intersection of separate indices on Cityname and Productname. Join indices can be used with bitmap representations for the RID lists for efficient join processing12.
Finally, decision support databases contain a significant amount of descriptive text and so indices to support text search are useful as well.
Materialized Views and their Usage
Many queries over data warehouses require summary data, and, therefore, use aggregates. Hence, in addition to indices, materializing summary data can help to accelerate many common queries. For example, in an investment environment, a large majority of the queries may be based on the performance of the most recent quarter and the current fiscal year. Having summary data on these parameters can significantly speed up query processing.
The challenges in exploiting materialized views are not unlike those in using indices: (a) identify the views to materialize,
(b) exploit the materialized views to answer queries, and (c) efficiently update the materialized views during load and refresh. The currently adopted industrial solutions to these problems consider materializing views that have a relatively simple structure. Such views consist of joins of the fact table with a subset of dimension tables (possibly after some selections on those dimensions), with the aggregation of one or more measures grouped by a set of attributes from the dimension tables. The structure of these views is a little more complex when the underlying schema is a snowflake.
Despite the restricted form, there is still a wide choice of views to materialize. The selection of views to materialize must take into account workload characteristics, the costs for incremental update, and upper bounds on storage requirements. Under simplifying assumptions, a greedy algorithm was shown to have good performance13. A related problem that underlies optimization as well as choice of
materialized views is that of estimating the effect of aggregation on the cardinality of the relations.
A simple, but extremely useful, strategy for using a materialized view is to use selection on the materialized view, or rollup on the materialized view by grouping and aggregating on additional columns. For example, assume that a materialized view contains the total sales by quarter for each product. This materialized view can be used to answer a query that requests the total sales of Levi’s jeans for the year by first applying the selection and then rolling up from quarters to years. It should be emphasized that the ability to do roll-up from a partially aggregated result, relies on algebraic properties of the aggregating functions (e.g., Sum can be rolled up, but some other statistical function may not be).
In general, there may be several candidate materialized views that can be used to answer a query. If a view V has the same set of dimensions as Q, if the selection clause in Q implies the selection clause in V, and if the group-by columns in V are a subset of the group-by columns in Q, then view V can act as a generator of Q. Given a set of materialized views M, a query Q, we can define a set of minimal generators M’ for Q (i.e., smallest set of generators such that all other generators generate some member of M’). There can be multiple minimal generators for a query. For example, given a query that asks for total sales of clothing in Washington State, the following two views are both generators: (a) total sales by each state for each product (b) total sales by each city for each category. The notion of minimal generators can be used by the optimizer to narrow the search for the appropriate materialized view to use. On the commercial side, HP Intelligent Warehouse pioneered the use of the minimal generators to answer queries. While we have defined the notion of a generator in a restricted way, the general problem of optimizing queries in the presence of multiple materialized views is more complex. In the special case of
Select-Project-Join queries, there has been some work in this area.14 15 16
Transformation of Complex SQL Queries
The problem of finding efficient techniques for processing complex queries has been of keen interest in query optimization. In a way, decision support systems provide a testing ground for some of the ideas that have been studied before. We will only summarize some of the key contributions.
There has been substantial work on “unnesting” complex SQL queries containing nested subqueries by translating them into single block SQL queries when certain syntactic restrictions are satisfied17 18 19 20. Another direction that has been pursued in optimizing nested subqueries is reducing the number of invocations and batching invocation of inner
subqueries by semi-join like techniques21 22. Likewise, the problem of flattening queries containing views has been a topic of interest. The case where participating views are SPJ queries is well understood. The problem is more complex when one or more of the views contain aggregation23. Naturally, this problem is closely related to the problem of commuting group-by and join operators. However, commuting group-by and join is applicable in the context of single block SQL queries as well.24 25 26 An overview of the field appears in a recent paper27.
Parallelism plays a significant role in processing massive databases. Teradata pioneered some of the key technology.
All major vendors of database management systems now offer data partitioning and parallel query processing technology. The article by Dewitt and Gray provides an overview of this area28 . One interesting technique relevant to the read-only environment of decision support systems is that of piggybacking scans requested by multiple queries (used in Redbrick). Piggybacking scan reduces the total work as well as response time by overlapping scans of multiple concurrent requests.
Server Architectures for Query Processing
Traditional relational servers were not geared towards the intelligent use of indices and other requirements for supporting multidimensional views of data. However, all relational DBMS vendors have now moved rapidly to support these additional requirements. In addition to the traditional relational servers, there are three other categories of servers that were developed specifically for decision support.
• Specialized SQL Servers: Redbrick is an example of this class of servers. The objective here is to provide advanced query language and query processing support for SQL queries over star and snowflake schemas in read-only environments.
• ROLAP Servers: These are intermediate servers that sit between a relational back end server (where the data in the warehouse is stored) and client front end tools. Microstrategy is an example of such servers. They extend traditional relational servers with specialized middleware to efficiently support multidimensional OLAP queries, and they typically optimize for specific back end relational servers. They identify the views that are to be materialized, rephrase given user queries in terms of the appropriate materialized views, and generate multi-statement SQL for the back end server. They also provide additional services such as scheduling of queries and resource assignment (e.g., to prevent runaway queries). There has also been a trend to tune the ROLAP servers for domain specific ROLAP tools. The main strength of ROLAP servers is that they exploit the scalability and the transactional features of relational
systems. However, intrinsic mismatches between OLAP- style querying and SQL (e.g., lack of sequential processing, column aggregation) can cause performance bottlenecks for OLAP servers.
• MOLAP Servers: These servers directly support the multidimensional view of data through a multidimensional storage engine. This makes it possible to implement front-end multidimensional queries on the storage layer through direct mapping. An example of such a server is Essbase (Arbor). Such an approach has the advantage of excellent indexing properties, but provides poor storage utilization, especially when the data set is sparse. Many MOLAP servers adopt a 2-level storage representation to adapt to sparse data sets and use compression extensively. In the two-level storage representation, a set of one or two dimensional subarrays that are likely to be dense are identified, through the use of design tools or by user input, and are represented in the array format. Then, the traditional indexing structure is used to index onto these “smaller” arrays. Many of the techniques that were devised for statistical databases appear to be relevant for MOLAP servers.
Several extensions to SQL that facilitate the expression and processing of OLAP queries have been proposed or implemented in extended relational servers. Some of these extensions are described below.
• Extended family of aggregate functions: These include support for rank and percentile (e.g., all products in the top 10 percentile or the top 10 products by total Sale) as well as support for a variety of functions used in financial analysis (mean, mode, median).
• Reporting Features: The reports produced for business analysis often requires aggregate features evaluated on a time window, e.g., moving average. In addition, it is important to be able to provide breakpoints and running totals. Redbrick’s SQL extensions provide such primitives.
• Multiple Group-By: Front end tools such as multidimensional spreadsheets require grouping by different sets of attributes. This can be simulated by a set of SQL statements that require scanning the same data set multiple times, but this can be inefficient. Recently, two new operators, Rollup and Cube, have been proposed to augment SQL to address this problem29. Thus, Rollup of the list of attributes (Product, Year, City ) over a data set results in answer sets with the following applications of group by: (a) group by (Product, Year, City) (b) group by (Product, Year), and (c) group by Product. On the other hand, given a list of k columns, the Cube operator provides a group-by for each of the 2k combinations of columns. Such multiple group-by operations can be executed efficiently by recognizing
commonalties among them30. Microsoft SQL Server supports Cube and Rollup.
• Comparisons: An article by Ralph Kimball and Kevin Strehlo provides an excellent overview of the deficiencies of SQL in being able to do comparisons that are common in the business world, e.g., compare the difference between the total projected sale and total actual sale by each quarter, where projected sale and actual sale are columns of a table31. A straightforward execution of such queries may require multiple sequential scans. The article provides some alternatives to better support comparisons. A recent research paper also addresses the question of how to do comparisons among aggregated values by extending SQL32.
7. Metadata and Warehouse Management
Since a data warehouse reflects the business model of an enterprise, an essential element of a warehousing architecture is metadata management. Many different kinds of metadata have to be managed. Administrative metadata includes all of the information necessary for setting up and using a warehouse: descriptions of the source databases, back-end and front-end tools; definitions of the warehouse schema, derived data, dimensions and hierarchies, predefined queries and reports; data mart locations and contents; physical organization such as data partitions; data extraction, cleaning, and transformation rules; data refresh and purging policies; and user profiles, user authorization and access control policies. Business metadata includes business terms and definitions, ownership of the data, and charging policies. Operational metadata includes information that is collected during the operation of the warehouse: the lineage of migrated and transformed data; the currency of data in the warehouse (active, archived or purged); and monitoring information such as usage statistics, error reports, and audit trails.
Often, a metadata repository is used to store and manage all the metadata associated with the warehouse. The repository enables the sharing of metadata among tools and processes for designing, setting up, using, operating, and administering a warehouse. Commercial examples include Platinum Repository and Prism Directory Manager.
Creating and managing a warehousing system is hard. Many different classes of tools are available to facilitate different aspects of the process described in Section 2. Development tools are used to design and edit schemas, views, scripts, rules, queries, and reports. Planning and analysis tools are used for what-if scenarios such as understanding the impact of schema changes or refresh rates, and for doing capacity planning. Warehouse management tools (e.g., HP Intelligent Warehouse Advisor, IBM Data Hub, Prism Warehouse Manager) are used for monitoring a warehouse, reporting
statistics and making suggestions to the administrator: usage of partitions and summary tables, query execution times, types and frequencies of drill downs or rollups, which users or groups request which data, peak and average workloads over time, exception reporting, detecting runaway queries, and other quality of service metrics. System and network management tools (e.g., HP OpenView, IBM NetView, Tivoli) are used to measure traffic between clients and servers, between warehouse servers and operational databases, and so on. Finally, only recently have workflow management tools been considered for managing the extract- scrub-transform-load-refresh process. The steps of the process can invoke appropriate scripts stored in the repository, and can be launched periodically, on demand, or when specified events occur. The workflow engine ensures successful completion of the process, persistently records the success or failure of each step, and provides failure recovery with partial roll back , retry, or roll forward.
8. Research Issues
We have described the substantial technical challenges in developing and deploying decision support systems. While many commercial products and services exist, there are still several interesting avenues for research. We will only touch on a few of these here.
Data cleaning is a problem that is reminiscent of heterogeneous data integration, a problem that has been studied for many years. But here the emphasis is on data inconsistencies instead of schema inconsistencies. Data cleaning, as we indicated, is also closely related to data mining, with the objective of suggesting possible inconsistencies.
The problem of physical design of data warehouses should rekindle interest in the well-known problems of index selection, data partitioning and the selection of materialized views. However, while revisiting these problems, it is important to recognize the special role played by aggregation. Decision support systems already provide the field of query optimization with increasing challenges in the traditional questions of selectivity estimation and cost-based algorithms that can exploit transformations without exploding the search space (there are plenty of transformations, but few reliable cost estimation techniques and few smart cost-based algorithms/search strategies to exploit them). Partitioning the functionality of the query engine between the middleware (e.g., ROLAP layer) and the back end server is also an interesting problem.
The management of data warehouses also presents new challenges. Detecting runaway queries, and managing and scheduling resources are problems that are important but have not been well solved. Some work has been done on the
logical correctness of incrementally updating materialized views, but the performance, scalability, and recoverability properties of these techniques have not been investigated. In particular, failure and checkpointing issues in load and refresh in the presence of many indices and materialized views needs further research. The adaptation and use of workflow technology might help, but this needs further investigation.
Some of these areas are being pursued by the research community33 34, but others have received only cursory attention, particularly in relationship to data warehousing.
We thank Goetz Graefe for his comments on the draft.
1 Inmon, W.H., Building the Data Warehouse. John Wiley, 1992.
3 Codd, E.F., S.B. Codd, C.T. Salley, “Providing OLAP (On-Line Analytical Processing) to User Analyst: An IT Mandate.” Available from Arbor Software’s web site http://www.arborsoft.com/OLAP.html.
5 Kimball, R. The Data Warehouse Toolkit. John Wiley, 1996.
6 Barclay, T., R. Barnes, J. Gray, P. Sundaresan, “Loading Databases using Dataflow Parallelism.” SIGMOD Record, Vol. 23, No. 4, Dec.1994.
7 Blakeley, J.A., N. Coburn, P. Larson. “Updating Derived Relations: Detecting Irrelevant and Autonomously Computable Updates.” ACM TODS, Vol.4, No. 3, 1989.
8 Gupta, A., I.S. Mumick, “Maintenance of Materialized Views: Problems, Techniques, and Applications.” Data Eng. Bulletin, Vol. 18, No. 2, June 1995.
9 Zhuge, Y., H. Garcia-Molina, J. Hammer, J. Widom, “View Maintenance in a Warehousing Environment, Proc. of SIGMOD Conf., 1995.
10 Roussopoulos, N., et al., “The Maryland ADMS Project: Views R Us.” Data Eng. Bulletin, Vol. 18, No.2, June 1995.
11 O’Neil P., Quass D. “Improved Query Performance with Variant Indices”, To appear in Proc. of SIGMOD Conf., 1997.
12 O’Neil P., Graefe G. “Multi-Table Joins through Bitmapped Join Indices” SIGMOD Record, Sep 1995.
13 Harinarayan V., Rajaraman A., Ullman J.D. ” Implementing Data Cubes Efficiently” Proc. of SIGMOD Conf., 1996.
14 Chaudhuri S., Krishnamurthy R., Potamianos S., Shim K. “Optimizing Queries with Materialized Views” Intl. Conference on Data Engineering, 1995.
15 Levy A., Mendelzon A., Sagiv Y. “Answering Queries Using Views” Proc. of PODS, 1995.
16 Yang H.Z., Larson P.A. “Query Transformations for PSJ Queries”, Proc. of VLDB, 1987.
17 Kim W. “On Optimizing a SQL-like Nested Query” ACM TODS, Sep 1982.
18 Ganski,R., Wong H.K.T., “Optimization of Nested SQL Queries Revisited ” Proc. of SIGMOD Conf., 1987.
19 Dayal, U., “Of Nests and Trees: A Unified Approach to Processing Queries that Contain Nested Subqueries, Aggregates and Quantifiers” Proc. VLDB Conf., 1987.
20 Murlaikrishna, “Improved Unnesting Algorithms for Join Aggregate SQL Queries” Proc. VLDB Conf., 1992.
21 Seshadri P., Pirahesh H., Leung T. “Complex Query Decorrelation” Intl. Conference on Data Engineering , 1996.
22 Mumick I.S., Pirahesh H. “Implementation of Magic Sets in Starburst” Proc.of SIGMOD Conf., 1994.
23 Chaudhuri S., Shim K. “Optimizing Queries with Aggregate Views”, Proc. of EDBT, 1996.
24 Chaudhuri S., Shim K. “Including Group By in Query Optimization”, Proc. of VLDB, 1994.
25 Yan P., Larson P.A. “Eager Aggregation and Lazy Aggregation”, Proc. of VLDB, 1995.
26 Gupta A., Harinarayan V., Quass D. “Aggregate-Query Processing in Data Warehouse Environments”, Proc. of VLDB, 1995.
27 Chaudhuri S., Shim K. “An Overview of Cost-based Optimization of Queries with Aggregates” IEEE Data Enginering Bulletin, Sep 1995.
28 Dewitt D.J., Gray J. “Parallel Database Systems: The Future of High Performance Database Systems” CACM, June 1992.
29 Gray J. et.al. “Data Cube: A Relational Aggregation Operator Generalizing Group-by, Cross-Tab and Sub Totals” Data Mining and Knowledge Discovery Journal, Vol 1, No 1, 1997.
30 Agrawal S. et.al. “On the Computation of Multidimensional Aggregates” Proc. of VLDB Conf., 1996.
31 Kimball R., Strehlo., “Why decision support fails and how to fix it”, reprinted in SIGMOD Record, 24(3), 1995.
32 Chatziantoniou D., Ross K. “Querying Multiple Features in Relational Databases” Proc. of VLDB Conf., 1996.
33 Widom, J. “Research Problems in Data Warehousing.” Proc. 4th Intl. CIKM Conf., 1995.
34 Wu, M-C., A.P. Buchmann. “Research Issues in Data Warehousing.” Submitted for publication.