Office of Learning and Information Technology
University of Wisconsin CIO Council
January 17, 2008
Regent Street, Madison, WI
- Oracle BI Insight Session Report
- Learn@UW Update
- Network Update
- Oracle Database Renewal
- Identity Management Procurement Update
- IT Reporting
- Records Management for Email
- Meeting dates
CIOs and their Representatives
Dave Schreiber from Oracle brought several of his colleagues to discuss business intelligence (BI). An Insight engagement was conducted with the UW System some six weeks ago, which generated a set of recommendations and a roadmap.
- A data warehouse is the main repository of an organization's historical data, i.e., its organizational memory.
- BI systems provide historical, current and predictive views of business operations based upon data from the warehouse, data marts, etc.
The Insight methodology includes:
- collecting information on the current state of affairs, in this case BI
- interviewing business and IT managers to understand requirements and processes to support key business objectives
- providing a roadmap to fully leverage the resources that are currently in place
- validating the priority of the initiatives based upon the organization's strategic plan and tactical initiatives
From the interviews, the UWS objectives for business intelligence appear to be:
- increasing operational efficiency
- enhancing data-driven decision making
- folding BI into the upcoming Human Capital Management (HCM) implementation
- pushing more BI capabilities out to end-users:
- understanding by IT of the key performance indicators needed for the consumers of the information
- understanding by consumers of the information that is available and how to access it
- creating a flexible, user-friendly, consolidated central data repository
As one of the interviewees said, "Technology is not our main concern."
The strengths that the UWS can leverage include:
- the timing of the HCM implementation
- support for BI from many functional areas
- experiences with previous data warehousing efforts
- existing expertise in several toolsets
Current versus future state
|Current State||Future State|
|Siloed and fragmented data||Unified, enterprise view|
|No single source of consistent, unambiguous data||Single data source, ensuring accurate and reconcilable results|
|Data modeled to optimize transaction capture and processing||Data modeled to optimize reporting and analysis by end users|
|Historical analysis||Capabilities for both historical and predictive analysis|
|Performance conflict between transaction processing and reporting||Reporting performance overhead offloaded to data warehouse where possible|
|Heavy reliance on IT and IR to interpret data structures and create reports||Intuitive data structures allowing users to create reports|
Objectives and Findings
Data Driven Decision Making
- Key Finding: Standardized and operational reporting is well-defined in some departments but no system-wide strategy exists for BI or analytical reporting.
- Encourage information sharing between departments and campuses
- Define key performance indicators (KPI) based upon the needs of functional managers and executives
- Define a system-wide BI and reporting strategy for HR and finance.
Data Governance and Stewardship
- Key Finding: Requirement to evaluate University system policies governing ownership, dissemination and use of data.
- Data stewardship at the data warehouse level must be established
- Policies governing sharing and use of data should be created and enforced; the leading higher ed institutions are underway with these efforts, but they are about at the same place as the UWS
- Security models should not hinder operational effectiveness and efficiency
- User training is necessary to ensure understanding of, and adherence to, data use policies
- Key Finding: Implementation of a system-wide data warehouse is critical to the success of the BI initiative
- Implement enterprise reporting and business intelligence tools along with the central data warehouse
- Implement monitoring technology to track usage and performance
- Eliminate shadow systems by providing system-managed data
Although the above sound fairly generic to all higher education institutions, that's because most institutions are about in the same place. However, the specific strategies, key performance indicators, etc. will vary from one institution to another.
From a two dimensional matrix of value versus existing/additive resources, the key initiatives, in recommended order of implementation, are:
1. Define system-wide HR and financial reporting strategies
1. Define key performance indicators
2. Establish data warehouse, data stewardship and data sharing policies
2. Evaluate security model changes
2. Encourage institutional information sharing
3. Implement enterprise data warehouse, BI and reporting tools
3. Eliminate shadow systems
3. Provide user training for data use policies
The above implementations can be started ASAP. The scopes of each initiative need to be defined, e.g., a data warehouse could be designed to do both analytical and standardized compliance reporting. Business practices and processes need to be defined before determining the KPIs. These might be tasks for the HR Working Group. Elena Pokot suggested using the feedback on reporting from the HCM Summit to help in the initial development of the KPIs. Lisa Wheeler suggested that it would also be possible to define KPIs from scratch by tapping into the expertise and needs of the provosts, CBOs and IR staff rather than just the IT and HR staff. It may not be possible to extract KPIs out of the thousands of current reports because many reports only exist because people have not had online access to the data. At smaller institutions, it might be possible to go to a few people and ask what are the few, critical reports they need.
Alan Foley, Jeanne Blochwitz and Peter Mann reviewed the performance of D2L at the end of last semester and discussed long term issues. Seven institutions, including UW Colleges, submitted e-grades throughout the semester for over 1,000 sessions and 316 instructors. This is a 75% increase over a year ago. The hardware has handled the increase well, but the D2L application is experiencing many timeouts. Cacheing needs to restart periodically, which increases the end of semester load.
To improve performance in the past year, there was an upgrade to version D2L 8.1. and two application servers were added. During the 3rd week of the December, nightly integration processes were suspended to extend the grade submission window. The database was indexed more frequently. Issues were escalated to the vendor via Kathy Pletcher and the Steering Committee.
Over the holidays, both D2L and Learn@UW Utility staff worked on performance issues. Compared with 2006, workloads showed a +40% increase in TCP connections with 900,000 logins during the 3rd week of fall 2007. Learn@UW is entering into a concentrated period with the D2L company to look at database performance issues. Action plans have been required from D2L on several outstanding issues that affect performance and Learn@UW is reviewing its internal response to incidents.
Learn@UW is looking into more sophisticated monitoring and quantification of incidents and identifying when there is a crisis that possibly warrants the risks of extreme corrective action.
John Krogman asked about potential impacts on performance due to the inability to archive old courses. Jeanne Blochwitz feels that the size of the database could be impacting performance however, in peak periods it is not the database that fails but the application that times out. The 8.3 release that the UWS intends to take in May 2008 provides the first version of a course purging, but not archiving, tool.
Elena Pokot inquired about how campuses can learn which of their course sections do e-grading and may therefore be affected by issues at Learn@UW. Learn@UW has statistics as to how many grades have been submitted per course.
Ken Splittgerber asked about input from network professionals regarding performance issues at the switch level. Peter Mann reported that network staff are engaged and there do not appear to be any bandwidth issues. The timeouts are purely within the application. David Lois explained that platform or backbone network issues should show up in multiple applications or institutions.
Chip Eckardt said that some of his faculty felt that the end-of-semester performance issues confirmed their predictions and perceptions of the systemwide utility and their faculty senate may recommend pulling out from Learn@UW to run their own instance. Kathy Pletcher has escalated these issues to D2L management.
In response to a question from Dave Dumke, Jeanne Blochwitz said that she didn't feel that expanded use of the gradebook was primarily responsible for the end-of-semester issues.
It is a goal to have someone from D2L be onsite at Learn@UW during the end-of-semester processing. Thus far, the vendor has resisted doing this, even virtually.
Brian Remer distributed the final bandwidth-based WiscNet billing numbers for Spring 2008. Base annual fees have been sent for the next couple years.
David Lois announced that Andrea Deau is leaving WiscNet. She had been focusing on education and her position will be filled after a regrouping.
BOREAS network services will be extended to North Dakota, but that is probably the limit.
The Badgernet Converged Network continues to be challenged by the ability of school districts and others to procure bandwidth at lower costs. DOA is renegotiating their contracts with bandwidth providers for lower rates. Institutions are also looking at applications that are much more bandwidth intensive than in the past.
Currently, TEACH subsidies are spread evenly regardless of need. This practice may be reconsidered, although changing it would be difficult.
Doug Flee and Lori Voss distributed a summary of the Oracle database license renewal negotiations. Contractual protections for price increases for maintenance on existing licenses expires in February 2008. The UWS is considering the licensing of additional products for security, high availability and leveraging investments in DBAs.
Real Application Clusters (RAC) is being considered for enterprise level unlimited licensing. RAC allows for running a single database instance on multiple machines which could result in hardware savings and protect against hardware failures. However, additional DBA training would be likely.
Active Data Guard would enable real-time synchronization of a standby database with an active database. The UWS already has the licensing to do non-real-time synchronization. Provisioning and Configuration Packs would enable DBAs to more efficiently patch and manage overall multi-database environments.
Current procurement authority for certain Oracle products is already covered by state procurement authority for sole sourcing. Not all Oracle products are covered by this authority, especially those that are considered to be in more competitive arenas.
John Krogman will join in subsequent discussions with Oracle.
Chris Holsman said that Tom Jordan from UW–Whitewater and the rest of the identity management procurement team did a fantastic job. The vendor proposals were both long and technical. A letter of intent to award an enterprise license for the UW System has been issued to Oracle for their identity and access management suite. A list of products was distributed and the first contract strategy meeting was held today. The pre-written connectors that interface to other applications are particularly critical. Training and professional help will be engaged for at least the high level IAA implementation.
The high level project objectives include:
- replacement of existing IAA infrastructure
- reengineering of upstream and downstream data connections
- supporting Peoplesoft HCM with full capability
- exploring opportunities for a cooperative identity management system for both enterprise and local use
A technical sandbox is being created for developers. UW-Madison will inventory the current IAA capabilities and assess gaps. Regional meetings or webinars will be held for identity and access management capabilities.
- DoIT for technical implementation, project management and communications
- IAM Task force for technical advice and policy recommendations
- IAA for system governance
- CIO Council
There is potential to expand local campus identity access infrastructures using these tools, perhaps through some sort of a FASTAR-like model. The delegated administrative capabilities of the system are quite robust.
A budget proposal is being made to the Common Systems Review Group for the initial license and implementation. The proposal is to replace all existing infrastructure and connections by the end of FY09.
At their February meeting, the Board of Regents (BOR) Business Finance and Audit Committee will determine the procedures for reporting on large and high risk projects to state government. The state has to approve only the board's policies on reporting. DOA and other state agencies have come up with dashboards for portraying the parts of projects.
For semi-annual reporting, the UWS institutions will report by December, since there isn't a January BOR meeting and the reports are due by March 1. Similarly, reporting will occur in June for the September 1 deadline since there often isn't a July or August BOR meeting.
For this year, the UWS institutions will send Ed Meachen their IT strategic plans such as they are. If an institution does not have a plan, a note should be sent to that effect. If an institution has a high cost, i.e., greater than $1M, or "vital" project, it will have to be reported separately. It is important that the Board of Regents not be blind sided. The definition of a "vital" project for the UWS and a UWS institution was discussed as well as the reporting and monitoring requirements for high cost and vital projects.
Laura Dunek and most of the members of the UW Records Officers Council joined the meeting.
Context on the UW Records Officers Council
The UW Records Officer Council was started a year ago as a result of a BOR policy document. The members of the council have, according to statute, been named by their chancellors to be responsible for records management at their institutions. The amount of time and the resources at their disposal vary by institution. They have monthly teleconferences. Their purposes are to:
- share resources
- distribute materials to UWS employees to assist them in managing records
- keep the UWS records management web resources up to date
Project for email e-records guidance
Since the advent of desktop PCs, the profession of records management has undergone a paradigm shift from paper to digital. There has not been a similar paradigm shift in applicable laws and regulations. The current project is to offer guidance on the basics of records management, particularly electronic and email records.
The two councils discussed several questions
What email system is utilized by your campus?
The majority of the institutions use Microsoft Exchange, but there was a wide diversity across the entire UWS including at least 70 email systems on the UW–Madison that don't report to DoIT.
What problems have the employees at your campus encountered when working with and managing the email contained within the system utilized by your campus?
- There are unintended consequences of what's going on behind the scenes with email systems and clients, e.g., some people may not be aware that their sent and "deleted" email is actually retained on their desktop and/or the server.
- Despite the blocking of more than 90% of incoming email, people are still receiving an overwhelming amount of email, which some tend to leave in their inboxes.
- Migration to new a email system can be traumatic for those who use it.
- Upgrades cause some problems for legacy email clients, e.g., some of the features of IMAP.
- Those who run the email system have no control over email that is backed up or archived outside of the system itself.
- Voice mail and email is commingled in "unified" boxes.
- It can be hard to trace the path of email messages going through the various servers.
- People don't appreciate limits on the sizes of attachments.
- People don't appreciate and struggle with mailbox quotas.
- People don't know how to effectively manage their email.
- There are problems with email delivery at some institutions.
- Volume of email that accumulates while one is away is daunting.
- Spam comes in both external and internal forms, e.g., campus announcements.
- The multiplicity of clients and devices for reading email results in management complexity.
- The forwarding of email to third-party accounts, even by faculty and staff, interferes with providing a reliable service.
- Significant resources are expended in filtering spam and tracking slow and lost email.
- Responding to records retention requests takes a lot of time.
- Spam filters catch false positive messages.
- People are using Blackberrys and other devices to receive email over cellular rather than IP networks.
- The push to the paperless office is driving up email volume and the need for additional resources.
- The lack of digital signatures makes email insecure.
- The lack of passwords on laptops that have email stored on them is a security risk.
- Some people leave their email in their inboxes rather than archiving it because the search tools work so well it doesn't pay to put messages into folders.
- Copying and blind copying email messages can obviate the responsible actions of others.
What email management solutions, resources, or tools are available on the campus email systems and what others would be useful?
Generally speaking, there aren't any tools for setting email message retention policies on a folder by folder basis. Consequently, people tend keep all of their email until they run up against quotas and then they delete something or archive a bunch of folders on a CD. There is little active records management taking place. People view their email as their personal property. At UW-Green Bay, people are encouraged to set up accounts for entities such as the Help Desk so that they don't feel that email belongs to them. Typically, messages that go into spam quarantine folders are automatically deleted on a time basis.
Do tools allow for automatic filing of messages?
Exchange and other email clients do allow for automatic filing, but actual use of the tools is a matter of training.
What about a UWS effort to develop records management tools?
This would be unlikely to to work effectively because of the variations across institutions. If retention requirements are known in advance, perhaps the capability of doing retention could be included in product acquisitions and acquisitions.
It is not well known that records management "clocks" don't necessarily begin on the date a record is created. The time limit could begin on the day a class ends or the day a person is hired.
A systemwide training effort on retention may be developed using desktop tools.
The next meeting of the UWS CIO Council will be on February 14 in Madison. Meeting dates, the directory of UW CIOs, and CIO Council meeting summaries are available at: www.uwsa.edu/olit/cio/.