You are currently browsing the tag archive for the ‘IBM’ tag.

In the weeks leading up to and as part of its Information On Demand Conference that my colleague assessed, IBM introduced version 8.5 of InfoSphere Information Server and several related product updates. As my colleague suggested earlier, IBM has an ambitious agenda to provide comprehensive information management capabilities through a combination of product development and acquisitions. The breadth of this portfolio is impressive, and InfoSphere Information Server 8.5 makes significant strides in tying the various pieces together.

In the same way that IBM has sought to unify capabilities further up the BI stack with IBM Cognos 10, as I assessed, one of the key pieces of InfoSphere 8.5 is a new interface. InfoSphere Blueprint Director provides a broader view of the process associated with data integration projects. It also improves reusability, access to metadata, life-cycle management and collaboration. At first blush Blueprint Director might look like a data integration workflow, but it is really much broader, extending to the reports and analytics for which the data is being prepared. Throughout the process you can access the specs and metadata using some of the same tools that end users access, such as the Business Glossary. It will be interesting to learn how far these blueprints extend the reusability paradigm. The key will be how fine-grained it is and how well the developers have encapsulated and exposed the end points, which could occur at any step in the process if it is granular enough.

This release includes many under-the-cover enhancements, too. Version 8.5 includes autodiscovery of warehouse schema and some of the business glossary information from existing systems. High availability via clustering and fail-over mechanisms is available in both production and development environments. The release also provides scalability and load balancing at the domain and database tiers for development teams. Parallel processing of XML data provides another area of performance improvements.

I was impressed to learn that IBM has created a shared transaction context from change-data capture all the way through the date integration and data quality stages so you can insure integrity and recoverability of the movement of data through the system. And support for third-party source-code control systems is available via the SCCS standard.

If your organization uses data in multiple languages, you can now apply data-quality rules to more locales, including Argentina, Brazil, Chile, Peru, Mexico, the Netherlands, India, Traditional Chinese and Japanese Kana.

In its data warehouse portfolio, IBM announced InfoSphere Warehouse enhancements, a new Warehouse Pack and extensions to its Smart Analytics Systems. Optim Performance Manager and IBM Mashup Center are now bundled with InfoSphere Warehouse. In addition, SAS functions can execute inside the DB2 database engine. SAS has been making efforts to partner with other data warehouse vendors to provide similar capabilities. It makes for interesting bedfellows in that IBM owns SPSS, which competes with SAS, but the collaboration is good for joint SAS/IBM customers because it minimizes the need to move data around in order to perform advanced analytics.

IBM also added a third Warehouse Pack to its offerings. Supply Chain Insight comes with a predefined data model and 20 Cognos reports designed to perform analyses across vendor performance, inventory and distribution aspects of the supply chain process. Supply chain analytics are generally well understood, as my colleague noted, so a packaged offering can automate and speed the process of implementation, but you should expect additional work to be required. IBM describes these analyses as “examples” of the types of analyses you would perform and provides customization guidelines.

Announced in April, IBM’s data warehouse appliance offering, Smart Analytics, has been extended with three new systems. Two systems based on x86 architectures have been added to the lower end of the product family, and a system using IBM’s Power technology has been added, including a version with solid-state storage. The result is a product line that is nearly as broad as Teradata’s but much less unified. It will be interesting to see how the Netezza acquisition will affect these appliance offerings, assuming the deal is completed. I would expect the Netezza products to play a prominent role in future Smart Analytics offerings, so we suggest caution with respect to purchasing the existing products.

In addition to the enhancements above, IBM updated a variety of other pieces of the stack. This is where I think IBM still faces some challenges. The product set is complicated and could use some rationalization. Several updates were announced. Initiate Master Data Service 9.5 has capabilities helpful for creating and managing trusted data with external data sources and partners. It also provides a mashup composer for operating on trusted data. Traceability Server 3.0 adds more connections to master-data sources, serial number management and performance enhancements. Data Architect 7.5.3 adds support for Cognos and InfoSphere Warehouse with autodiscovery capabilities. And Guardiam Data Redaction 1.1.2 supports French, German and Spanish and can provide redaction for unstructured documents in FileNet.

Somewhat buried in these announcement was a technology preview of Hadoop-based “big data” analytics software called IBM InfoSphere BigInsights running both on premises and in the IBM Test Development cloud. It is only available for development purposes currently, but it is worth noting because it indicates the importance of Hadoop to IBM, which has even created its own distribution of Hadoop. I expect to see other vendors embracing Hadoop as well, as indicated in my earlier blog post regarding Hadoop World.

IBM is a force to be reckoned with in information management. It competes head-to-head with Oracle and SAP and also with Informatica, which has been steadily building its portfolio through acquisitions. IBM’s recent announcements help unify its stack. In contrast with the Cognos 10 portion of the stack I assessed recently, the company seems to have done a better job of integrating with third-party technologies but still has further to go in unifying the many components it offers. That said, the integration with its offerings for master data management and data warehousing make IBM InfoSphere Information Server 8.5 worthy of consideration.

Let me know your thoughts or come and collaborate with me on  Facebook, LinkedIn and  Twitter .


David Menninger – VP & Research Director

On October 25, IBM introduced Cognos 10 at its Information on Demand and Business Analytics Forum in Las Vegas that I attended to review the technology closer from my examination at its recent IBM Business Analytics analyst summit in September. According to Rob Ashe, IBM’s general manager of business analytics, Cognos 10 has been developed for over six years. You’re probably aware that in that period IBM made a variety of acquisitions including Cognos itself. These acquisitions and their impact on the new product are clearly in evidence as part of the release.

Overall, Cognos 10 provides a unified user interface that integrates a broader set of analytics and technology as a result of the acquisitions of SPSS, Applix TM1 and Celequest. More than half the respondents (57%) in our recent benchmark research indicated that providing access through a variety of tools was an important goal. Cognos 10 also introduces a common workspace for all the different analytics that are now supported. Company officials described the changes as moving from role-centric capabilities to a unified workspace with “progressively more interaction” based on the user’s knowledge, interests and job function. In practice it’s a transformation from product-based capabilities to the common workspace because each type of analysis and capability was, in the past, based on a separate product: reporting, ad hoc query, planning, statistics, real-time monitoring and others.

Progressive interaction is a powerful technique that can greatly enhance the range of users and use cases that a single product can support. I attribute the popularity of spreadsheets, in part, to this design concept. Novice users can read a worksheet with almost no prior familiarity. With a small amount of training, users can create spreadsheet formulas. More advanced users can create nearly pixel-perfect reports with embedded graphics. And high-powered “spreadsheet jockeys” can use macros and scripting languages to create something close to an entire application with user input, data validation and integration with other elements of an IT infrastructure. Cognos 10 attempts to create a similarly broad and seamless range of capabilities, and for the most part it succeeds.

Why did IBM skip version 9 and go directly from Cognos 8.4 to Cognos 10? In part it was because this would be a big release, but as well because IBM had acquired products that were already on a 9.x version such as TM1. I agree with that thinking but not with the decision to number this release 10.1. It’s a gimmick to make you think it is not dot-zero release and thus won’t suffer from the quality issues common in dot-zero releases. That said, I have no reason to question the quality of the release. In fact, there are specific life-cycle management features designed to support the upgrade process, including testing of new and old reports side-by-side, but I’d rather see the emphasis on touting these features and the quality of the release than using a version-number gimmick.

And Cognos 10 has plenty of new features to talk about. Search, like other capabilities, has been brought into the common workspace and is now simply another available feature rather than a separate product. It is well integrated with the entire stack including single search capabilities across data, metadata (including data lineage), analyses and collaboration about the decision-making process. However, search is still focused primarily on structured data and related documents.

Mathematical models, statistics and predictive analytics are readily available in the new version. The mathematical models common to financial performance management applications are key to enabling users to project the operations of their organization into the future. Business intelligence need not be just about analyzing history. Our benchmark research indicates 84% rank planning and forecasting as very important or important characteristics to add to BI and fulfills on a common platform for also supporting what our firm calls and has researched with integrated business planning. Statistics and predictive models are also useful for predicting the outcomes of future activities and individual transactions rather than the future of the overall organization.
This release also includes self-contained interactive content in what Cognos calls “active reports.” If, for example, you are on a plane or at a customer site where you can’t be connected, active reports allow you to continue to investigate a portion of the data that has been associated with the report. And since Cognos 10 has write-back and what-if capabilities, the changes will be synchronized when you reconnect.

The mobile part of Cognos 10 is not ready for release yet. Presenters gave early demonstrations of it, but these capabilities will have to be fully assessed and compared to other offerings later. Mobile BI, as my colleague has assessed, is changing as business demands it and IT starts to embrace mobility beyond a single corporate standard.

One of the areas where Cognos is breaking down barriers outside of the traditional BI stack is in the area of collaboration. As an industry, vendors have not invested enough in bringing BI to users – they typically ask the users to come to BI. (See my thoughts on this issue <> for more.) Cognos 10 recognizes that BI is part of a decision-making process and includes collaborative capabilities to support that process. Lotus Connections is the backbone for Cognos 10 collaboration and is included as part of the product. Using it, decisions can be noted and followed through to completion. Links to e-mail provide alerts to participants in the process. They can view the backgrounds of others on the decision-making team. And in a world of increasing compliance requirements, the decision-making process can be reviewed after the fact if necessary. These are steps in the right direction, and I’d like to see it go even further with tight integration to non-IBM collaboration environments and productivity tools.

As with most major releases, this one includes performance enhancements. They shorten the time for both cube-building and querying. Cognos 10 introduces a feature called Dynamic Query which leverages a caching scheme and query-routing techniques to speed the queries by directing the queries to the cache (if available) or to an OLAP cube or relational source as appropriate. Dynamic query applies to non-Cognos cubes as well as Cognos cubes.

Life-cycle management is a favorite feature of anyone who has ever managed a large BI upgrade. Cognos 10 has multiversion capabilities that allow organizations to run a mixed environment of old and new reports, so you can transition gradually to the new version. In addition, you can compare the output of old and new reports to highlight differences directly within Lifecycle Manager. This feature should save IT organizations significant amounts of resources during the testing process.

At the launch event, Cognos had customers from Boeing onstage talking about their upgrade process. Their company runs 13 different applications based on Cognos for 3,500 users. The presenters said they were able to complete the upgrade in six days. They installed it in their development environment on a Monday, in their test environment on Tuesday, tested it from Wednesday to Friday and made a go or no-go decision on Friday afternoon. On Saturday they installed it in production, and when users came to work on Monday they were all on the new system. While every environment will be different, that speed of implementation seemed pretty impressive to me.

However, there’s still more work to be done to incorporate an even broader range of business users and more types of analyses. Specifically, I’d like to see some of the Cognos 10 features integrated more with a range of non-IBM products including collaboration tools and metadata. As well I’d like to see advanced analytics, still primarily the domain of power users, made more widely usable. Cognos has made strides in delivering advanced analyses to a range of users, but developing advanced analytics still requires specialized skills. I’d also like to see better incorporation of unstructured data analysis. The SPSS text analytics capabilities could be surfaced more and search could be expanded to deal better with the combination of structured and unstructured as well as internal and external data.

Overall, Cognos 10 is a dramatically improved release and in my opinion currently sets the standard for end-to-end integrated business intelligence products and then supporting analytics, planning and performance management. I expect it also creates an easier platform for incremental developments on the part of the IBM Cognos development team going forward, since it has tackled many of the hardest integration challenges already. IBM has clearly communicated its intention to make more acquisitions, with another $14 billion earmarked for that across the company, so we can expect the new platform to be expanded even further in the future.

Let me know your thoughts or come and collaborate with me on  Facebook, LinkedIn and  Twitter .


David Menninger – VP & Research Director

Follow on

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 17 other followers

David Menninger – Twitter

Ventana Research

Top Rated

Blog Stats

  • 43,449 hits

Get every new post delivered to your Inbox.

%d bloggers like this: