IBM Cognos – Dynamic Cubes - Northern California Cognos · PDF file · 2015-05-13Current Month All Sales cube All Sales Current Month Sales Historic ... IBM Cognos –...

© 2015 IBM Corporation

IBM Cognos – Dynamic Cubes


Scope of Discussion

General Architecture

IBM OLAP technologies

2


Elevator Pitch

In-memory OLAP cubes

High performance dimensional analytics over growing data volumes

Relational sources with star or snowflake schemas

Aggregate aware: database and/or memory based aggregates

Built-in aggregate optimization

Extends DQM in-memory caching of members, data, expressions, results, aggregates

Accessible by all IBM Cognos Interfaces

Included with Cognos Business Intelligence (no additional cost)


ARCHITECTURE

Dynamic Cubes / IBM Cognos BI


A Basic Conceptual View of a BI Stack

Presentation Layer

Application Layer

Data Layer


Presentation Layer

Application Layer

Data Layer

… slightly more detail.

Aggregate

LayerSemantic Layer


Cognos Architecture


Modern and Legacy

Sources

Application

Sources

3rd Party OLAP

Sources

Relational

Sources

Dynamic

Query

Mode

Common Business Model

Classic

Query

Mode

Scorecards

Dashboards

Reports

Ad-hoc

Query

Analysis

& Exploration

Trend &

Statistical

Analysis

What-If

Analysis

PowerCubes

Open Data Access

OLAP

Over

Relational

Dimensionally

Modeled

Relational

Large Enterprise Data Warehouse

Database

Aggregates

Dynamic

Cubes

TM1


– Caches are

generally

shared across

all users

– Security is

applied on top

of the caches

Dynamic Query

Database

Data Warehouse

Database Aggregates

Result Set Cache

Expression Cache

Member Cache

Data Cache

Aggregate Cache

Database

Data Warehouse

Database Aggregates

Database

Data Warehouse

Data Warehouse

Extensive Caching


Details of the various caches

Member Cache - The hierarchies defined in a

cube model are all loaded into memory when the

cube starts.

Aggregate Cache - the cube will retain the

aggregate values in separate cubelets in a

separate cache, up to the amount of space

specified in the configuration for the cube.

Data Cache - Any queries that retrieve data from

the underlying database, or further aggregate

data from the aggregate cache, are stored as

cubelets in a separate cache.

Result Set Cache - the result set of each MDX

query executed by the engine is stored within the

on-disk result set cache.

The Expression Cache is used by the MDX

engine to cache the result of set expressions that

operate on large sets and output a much smaller

set for subsequent reuse, either within the same

query or within subsequent MDX queries

processed by the MDX engine.

Extensive Caching


Virtual cube used as source for another

virtual cube

Combines cubes with common Time dimension

Virtual cubes combine two

cubes

Combines cubes with nearly identical

dimensions

Inventory

Sales

SalesInventory

Store

Sales

Web

Sales

Virtual Cubes

11


TimeCurrent Month

All Sales cube

All

Sales

Current

Month

Sales

Historic

Sales

Virtual Cubes: Low Latency

12


THE DC CYCLE

Soup to Nuts


Interfaces Used

Developer/Administrator

Cube Designer

Cognos Administration Console

Dynamic Query Analyzer

End User

Workspace / Workspace Advanced

Any other


1. Model & publish

2. Deploy & manage3. Reporting & analytics

4. Optimize

Dynamic Cube Server (App Server)

DynamicCube

Logs

CM

Warehouse

Dynamic Cubes Lifecycle

15


Cube Designer


Cognos Administration


Easily generate aggregates to optimize

performance

Based on model and/or workload

(report, package, user, or time)

Advisor recommends:

– In-memory aggregates

Based on data heuristics

Based on user-defined aggregates

Loaded into cache on startup

– In-database aggregate tables

Detailed SQL script for DBA

Update model and redeploy cube

Save history of recommendations for later reuse

18

Dynamic Query Analyzer / Aggregate Advisor


Advisor Recommendations

19


Dynamic Query Analyzer / Aggregate Advisor

© 2015 IBM Corporation21

Workspace Advanced


PARITY & NON-PARITY

PowerCubes & Dynamic Cubes

© 2015 IBM Corporation23

Dynamic Cubes PowerCubes TM1

Extensive in-memory caching

for performance

File-based In-memory technology Write-

back support

Optimal for read-only reporting

and analytics

Interactive analysis to large

number of users

Optimal for

Write-back

What-if Analysis, Planning &

Budgeting

Other Specialized applications

Star or snowflake schema

required

Flexible data source

requirements

Dimensional data source

optimal.

Star schema not required

Supports in-memory, in-

database and on-demand

aggregation

File-based cube with 100% pre-

aggregation

Aggregation occurs on demand

Multiple terabytes File-based limits (2GB) Medium data volumes


Dynamic Cubes Hardware Requirements

24

Configuration* Small Medium Large Extra Large

Description Development, POC,

small business app

Medium

enterprise app

Large

enterprise app

SKU level data

# members 600,000 3,000,000 15,000,000 30,000,000

# users 1-100 100-1,000 1000-5,000 5000-10,000

CPU Cores 4 4-8 8-16 16-32

Memory 3 GB 7 GB 45 GB 135 GB

Disk 1 GB 10 GB 50 GB 100 GB

(*) Configurations based on version 10.2.1 FP1 or later.

(*) Cube with 12 measures and 12 dimensions


Dynamic Cubes - Summary

High Performance

– 80x improvement with aggregates*

– 80% queries under 3 seconds*

– Over 50% queries sub-second*

Growing Data Volumes

– In the labs testing with terabytes of

data

Flexible and Optimized

– Aggregate Advisor to easily create

optimized aggregates

Maximize Value of Data Warehouse

– Aggregate awareness to balance

load across app and DB tiers

25(*) Based on benchmarks conducted on IBM Labs.


Demonstration:

26

• Explore the Cube Designer interface

• Explore dimension properties

• Explore cube properties

• Publish cube to Cognos Connection

• Use Dynamic cube in a report

IBM Cognos – Dynamic Cubes - Northern California Cognos · PDF file · 2015-05-13Current Month All Sales cube All Sales Current Month Sales Historic ... IBM Cognos –...

Documents