Big Data Analytics with IBM Cognos Dynamic Cubes

IBM Redbooks Solution Guide


With IBM® Cognos® Business Intelligence, you have a proven enterprise business intelligence (BI) platform with an open-data access strategy. Because Cognos Business Intelligence uses a business language that is relevant to consumers, you can pull data from various data sources. Then, you can package it into a business model and make it available to consumers in various interfaces that are suited to the task.

IBM Cognos Dynamic Cubes, which is a feature of the Cognos Business Intelligence V10.2 software, complements the existing query engine. As explained in this IBM Redbooks® Solution Guide, it extends Cognos scalability to enable speed-of-thought analytics over terabytes of enterprise data, without being forced to rely on a new data-warehousing appliance. With this capability, which adds a new level of query intelligence, you can unleash the power of your large enterprise data warehouse.

For related information about this topic, refer to the following IBM Redbooks publication:
IBM Cognos Dynamic Cubes, SG24-8064-00


IBM® Cognos® Dynamic Cubes, which is a feature of the Cognos Business Intelligence V10.2 software, complements the existing query engine. It extends Cognos scalability to enable speed-of-thought analytics over terabytes of enterprise data, without being forced to rely on a new data-warehousing appliance. With this capability, which adds a level of query intelligence, you can unleash the power of your large enterprise data warehouse. Figure 1 illustrates how the Cognos Dynamic Cubes feature is integrated into the Cognos Business Intelligence stack.

Cognos Dynamic Cubes integrated into the Cognos Business Intelligence stack
Figure 1. Cognos Dynamic Cubes integrated into the Cognos Business Intelligence stack

Did you know?

In 2012, about 2.5 quintillion bytes of data were created every day. Social media is now generating petabytes of data per day.

Business value

With social data generating petabytes per day, and instrumented devices becoming the norm, data volume growth is accelerating at an unprecedented pace. Big data is a growing business trend, with data from unconventional sources having the potential to be business disruptors. However, before the power of these new sources can be fully used, you must understand what is happening within your own business. Understanding your own business is added value of a data warehouse and is why taking full advantage of these data holdings is a critical first step to using these new sources of data. In addition, any organization that relies on instrumented infrastructures can maximize the efficiency of its operations. Analytics is key to accomplishing this type of optimization, leading to concrete business results.

Data warehouses are the recognized foundation for enterprise analytics. By using data warehouses, an organization can bring together cleansed data from separate sources of input, both internal and external, such as from partners or suppliers. Instead of garbage in, garbage out information to support decision-making, a consistent and consolidated enterprise-wide view of data from a business provides the foundation to improve your business. Building upon a trusted information platform for analytics is a key contributor to long-term business health. Not only do data warehouses enable higher quality information, they enable high-performance data access for analytic-style applications. IBM Cognos Dynamic Cubes technology helps to use the core strengths of an enterprise data warehouse and take it to the next level of performance for analytics, making the deploying and tuning easier and faster.

Solution overview

The IBM Cognos Dynamic Cubes technology is meant to solve a specific but growing business problem, enabling high-performance interactive analysis over terabytes of data in an enterprise data warehouse. As data volumes grow, analyzing that data with speed-of-thought performance can be challenging. Even with modern data warehouse technology, some operations require significant computation or data movement. This computation or movement creates delays and reduces the satisfaction of business users who want to perform these analyses.

Various ways exist to accomplish performance over large volumes of data. From self-contained cubes to large in-memory appliances, different vendors are employing variations of similar methodologies to give business users timely response times. The Cognos Dynamic Cubes technology aims to give maximum flexibility in how memory is used to accelerate interactive analysis over terabytes of data so that you can evolve your deployments over time.

Solution architecture

The IBM Cognos Dynamic Cubes technology is part of the IBM Cognos Business Intelligence query stack and is available with existing IBM Cognos entitlements. It provides a new and powerful tool to enable high performance analytics over large data warehouses (see Figure 2).

A high level representation of the Cognos Dynamic Cubes architecture
Figure 2. A high-level representation of the Cognos Dynamic Cubes architecture

The Cognos Dynamic Cubes solution consists of IBM Cognos Cube Designer (a modeling tool ), a dynamic cube object in the administration environment (which becomes the data source), and Aggregate Advisor (a wizard) that is started from Dynamic Query Analyzer:
  • IBM Cognos Cube Designer

    Cognos Cube Designer is a modeling tool that brings together the best modeling principles from past successful modeling technology, with a modern and extensible architecture. The first step to deploying Cognos Dynamic Cubes is to model with the Cognos Cube Designer.

  • IBM Cognos Dynamic Cubes Server

    After a dynamic cube is designed and deployed, it becomes available in the Cognos environment and acts as the data source to the interface layer for dynamic cube applications. It manages all aspects of data retrieval and leverages memory to maximize responsiveness, giving you full flexibility to manage what is in memory and when you want to refresh in-memory data. You manage dynamic cubes in the Cognos administration environment.

    A dynamic cube contains several in-memory elements to drive performance:

    • Metadata members
    • Aggregates
    • Data
    • Results sets
    • Expressions

  • Aggregate Advisor (part of IBM Cognos Dynamic Query Analyzer)

    The Aggregate Advisor scans cube definitions and usage logs. It also recommends aggregates to improve performance. This approach helps to more easily address specific performance problems.

Usage scenarios

The Cognos Dynamic Cubes solution applies to the following usage scenarios.

Multigrain fact scenarios

A common modeling problem is multigranularity, which occurs when levels of dimensional detail for facts differ. In your data source, the level of information in a dimension can be more precise than the fact data of some fact tables in which the dimension takes part. For example, a time dimension can have dimension information for the levels of year, quarter, month, and day. For a sales fact table, the facts exist at the day level. For a fact table with planned sales values, the fact grain or level is probably at a higher level of detail, such as month. This difference in fact grain can make it difficult to plan queries correctly if a report user included a level below the fact grain in the report.

The method for handling this scenario is to create role-playing dimensions for each instance of a dimension that has varying levels of granularity to different fact tables. In this example, you need to model a time dimension down to the day level for the sales cube and to model a time dimension down to the month level for the sales target cube. For the day level, the relationship is formed between the appropriate keys in the fact table and the day level. For the month level, the relationship is formed between the month level and the fact table keys.

If you want to create reports that used data from both cubes, you can create a virtual cube. The virtual cube can have the two time dimensions merged. The levels of both source cubes will be merged. The virtual cube might have levels that are below the grain of measures from one of the source cubes. Queries that are made in the virtual cube that use dimension levels below the grain of a measure will return null values, ensuring that the consumers of the cube do not have double counting.

Perhaps you have the following cubes:
  • A sales fact cube, where the fact grain for the time dimension is at the day level
  • A planned sales cube, where the fact grain for the time dimension is at the month level

By using a virtual cube that employs both of these cubes as its source, you can make a query with day-level objects (or a member of that level, depending on the studio that you are using) against the sales facts. Then, you get the results and the expected null values for the planned sales facts. If you used a time dimension grain that was common to both fact tables, you get non-null values for measures from both fact tables.

Aggregate cubes

A second scenario addresses situations where a higher level of granularity is desirable through aggregate awareness. This awareness is accomplished through aggregate cubes, which are supported by IBM Cognos Dynamic Cubes. Aggregate cubes define the measures, dimensions, and dimension grain by which queries can be routed to aggregate tables rather than to the detail fact table. Because aggregate tables store fact data at a higher-than-detail level of granularity, the time necessary to aggregate values during the query can be lessened, improving performance. A query can be routed to the aggregate table if all the measures and dimension hierarchies of the query exist in the aggregate cube definition. Not all of the dimensions and measures in the aggregate cube need to be in the query.

The objective of modeling an aggregate cube is to establish rules by which the dynamic cube can detect when it can route a query to an aggregate table. This task is done by specifying a mapping from the identifiers in the dimensions and measures in the cube that have scope to the aggregate table, to the identifiers in the aggregate table. If necessary, mapping is done to its related tables in a rolled-up dimension schema.

This aggregate cube routing directs a query only to the aggregate table for a query that uses objects from a dimension grain at or above the grain of the mapping between it and the aggregate table. Therefore, using objects from a grain below the mapping grain does not cause double-counting because that query continues to route to the detail fact table.


IBM Cognos Dynamic Cubes is tightly integrated into the Cognos Business Intelligence stack, and its data can be surfaced through any of the Cognos interfaces. With this method, existing customers can integrate this technology into their application environment without affecting existing users. Such users are already familiar with interfaces such as Report Studio, Business Workspace, and Business Workspace Advanced (previously named Business Insight and Business Insight Advanced).

Different data requirements require different data solutions. One data path cannot be proficient at solving widely different data problems. Therefore, IBM Cognos has technologies that are built to suit specific application requirements. Table 1 can help you better understand the primary use case for each technology. However, carefully consider your individual application requirements when you make such a decision.

Table 1. Use cases for Cognos Dynamic Cubes technology
Cube technologyPrimary use cases
IBM Cognos TM1®, in-memory cube technology with write-back support
  • It is optimal for write-back, what-if analysis, planning and budgeting, or other specialized applications.
  • It can handle medium data volumes. The cube is run 100% in memory.
  • Aggregation occurs on demand, which can affect performance with high data and high user volumes.
Dynamic Cubes, in-memory accelerator for dimensional analysis
  • It is optimal for read-only reporting and analytics over large data volumes.
  • It provides extensive in-memory caching for performance, backed by aggregate awareness to use the power and scalability of a relational database.
  • A star or snowflake schema is required in the underlying database (used to maximize performance).
PowerCubes, file-based cube with pre-aggregation
  • It is optimal to provide consistent interactive analysis experience to many users when the data source is an operational or transactional system, and a star or snowflake data structure cannot be achieved.
  • The pre-aggregated cube architecture requires careful management, by using cube groups to achieve scalability.
  • Data latency is inherent with pre-aggregated cube technology, where data movement into the cube is required.
OLAP Over Relational (OOR), dimensional view of a relational database
  • It is optimal to easily create a dimensional data exploration experience over low data volumes in an operational or transactional system, and where latency must be carefully managed.
  • Caching on the Dynamic Query server helps performance.
  • Processing that is associated with operational or transactional system affects performance.

Supported platforms

For information about software environments that are supported in IBM Cognos Business Intelligence V10.2, see "Cognos BI 10.2 Software Environments" at:

Ordering information

Table 2 lists the programs that IBM Cognos Business Intelligence contains.

Table 2. Program numbers and names
Program numberProgram name
5724-W12IBM Cognos Business Intelligence V10.2.0
5724-W68IBM Cognos Business Intelligence PowerPlay® V10.2.0
5724-W19IBM Cognos Business Intelligence Reporting V10.2.0
5724-W13IBM Cognos Data Manager V10.2.0
5724-W20IBM Cognos Mobile V10.1.0

Related information

For more information, see the following documents:

Special Notices

This material has not been submitted to any formal IBM test and is published AS IS. It has not been the subject of rigorous review. IBM assumes no responsibility for its accuracy or completeness. The use of this information or the implementation of any of these techniques is a client responsibility and depends upon the client's ability to evaluate and integrate them into the client's operational environment.


Publish Date
07 December 2012

Rating: Not yet rated


IBM Form Number