Patrick Ellis, BTS—Enterprise Systems Solutions
Peter Cava, IST—Data Services
There's a lot happening with the data warehouse, starting with an upgrade to the Hyperion Business Intelligence toolset used by users, but going far beyond that. Among the other initiatives now getting underway are the selection and implementation of a metadata management tool; redesign of the tables, views, and metadata topics of the data warehouse to integrate data from multiple major campus systems; and incorporation of more key operational data sets. The goal of this multiyear strategy is to turn the data warehouse into a true enterprise data warehouse, with broad and integrated sets of consistent, reliable, easily accessible, and secure data.
At the end of May 2008, the software upgrade of the BAIRS portal to Hyperion System 9.3.1 will be released to the campus. This upgrade will at first be nearly a "pure technical" upgrade, with few new features released at that time. However, new functionality not included in the upgrade rollout will be released at later dates. The pressing reason for this upgrade is that the current implemented version of the tool goes out of vendor support in September, hence the drive to get it done quickly and not burden the upgrade with delivering new functionality.
Nonetheless, there will be some immediate benefits. The most obvious immediate benefit of the upgrade for many users and their desktop support staff will be broader operating system and browser support by elimination of the requirement to use Internet Explorer with an installed plugin. This implementation is known as the "Zero Client" and will be suitable for those users who use the prebuilt dashboard reports and do little or no modification to the queries "behind the dashboard". XP, Vista, and Mac OS X are all supported for the Zero Client. Browser testing is underway, and it appears that IE6, IE7, Firefox 2, and Firefox 3 may be supported, with Firefox usable on either Windows or Mac OS. Safari tests are so far not encouraging, however.
For more advanced users who require greater "behind the dashboard" query control, there will be an implementation called "Interactive Client", with an IE6 and IE7 plugin needed. There is, regretfully, still not a plugin for any Mac OS compatible browser, but for those advanced users on Macs, access through the Citrix client will remain available.
Other important differences delivered with the upgrade will be online user folders for saving user-customized reports, and a much more robust and fail-safe architecture. Features being studied for possible rollout later, if resources permit, include Hyperion integration with Excel, Word, and PowerPoint; online sharing of user-customized reports; email alerts of data updates and other system events; user-scheduled automated reports; user-creation of monitoring boards with graphic presentations of key operations parameters; drill-down capabilities; and the capacity to merge data outside the data warehouse with those in the data warehouse using the Hyperion toolset.
Due to the proximity of fiscal close to the rollout, we will continue to run the current Hyperion 8.2.1 version in parallel with the new Hyperion System 9.3.1 version through fiscal close, with both accessing the same production reports, schemas, and data. This is being done to allow users to choose to defer switching to the new version at a time when their data demands are heavy. Once the fiscal year has closed, we will shut down access to the 8.2.1 version.
To provide an infrastructure for the unified management of metadata across the campus, the purchase and implementation of a metadata management tool has been funded. Market analysis is now being done to gain in-depth understanding of the tools on the market, and we are about to form a cross-campus advisory group to help define requirements for the tool. Currently, metadata is not unified across the many campus datasets, which leads to conflicts in data that result in misleading or even conflicting data being used by campus decision makers. Once implemented, this tool will give us and other proprietors of large datasets a single repository of both technical and functional metadata, where a complete view of campus data can be had, providing a base for establishing a clear and consistent understanding of the data.
Another initiative just starting is a redesign of the core data dimensions used in the enterprise data warehouse. Currently, dimensions that define such basic things as timescales and identity are different for each dataset in the warehouse, making it difficult to integrate data from BFS, HRMS, and PPS, let alone from sources currently outside the data warehouse. The new campus data architect, Greg Hamilton, is starting to build new dimensions to not only reflect the needs from the data currently in the data warehouse, but also those of data from other systems, including student, donor, facilities, and other operational systems. These will be intended not only for use by data in the data warehouse now and in the future, but also so they can be adopted for data outside the enterprise data warehouse to provide a consistent form to campus data.
This work is the first step in a long-term project to resolve and document metadata and taxonomy alignment, and the differences between all major campus operational systems; map data flows to determine and document the source systems and other systems holding data; and design a common metadata schema to ensure accurate, consistent, and flexible analysis of operational data. This will be a difficult but essential task, requiring assistance and collaboration from system owners and data proprietors across the campus to untangle the existing data discrepancies and transformations between systems that were developed independently of one another and yet now must operate as a single, unified system. The payoff will be campuswide data under a single schema that is consistent and understandable, making data analysis and data-driven decisions more feasible, efficient, accurate, and effective.
Finally, we are talking with numerous campus partners about hosting their data in the data warehouse, and, in some cases, making at least some of that data accessible to others on campus to aid in planning and analysis. These talks are in the early stages, and are not yet to a point where we can say which data may or may not end up in the data warehouse for use by others.
[ iNews | Search | IST | UC Berkeley Computing | UC Berkeley ]
iNews: UC Berkeley information technology news channels
Copyright 2008, The Regents of the University of California