Office of Learning and Information Technology

UWS CIO Council Meeting

December 10, 2004 – Pyle Center, Madison, WI

Madison, WI

Questions to david@uwm.edu

Attendees

CIOs and their Representatives

Guests

Mark Anderson
John Berens
Joe Brown
Nancy Crabb
David Dumke
Jack Duwe
Ron Kraemer
John Krogman
Jim Lowe
Bruce Maas
Ed Meachen
Kathy Pletcher>
Elena Pokot
Carrie Regenstein
Gary Smith
David Stack
Annie Stunden
John Tillman

Gail Bergman
Perry Brunelli
Lorie Docken
Ken Ebbe
David Hart
Melissa Kepner
David Lois
Jody McCann
Sue Michalek
Mike Mietz
Leslie Perelman
Brian Remer
Michael Schlicht
Tom Scott
Bob Stuessy
David Wirth

BadgerNet Converged Network (Mike Mietz, Jody McCann and Bob Stuessy)

Efficiently meeting the needs of the UW System is a top priority of the BadgerNet Converged Network initiative. Not every development can be communicated for a couple weeks, because contract negotiations are underway. A full communication process will then go into effect.

In 2001, the expiration of BadgerNet video and data contracts were on the horizon. An analysis aimed at meeting the needs of education in 2010 and beyond included meetings, focus groups, a legislative business case and a consideration of governmental applications. Plans and directions got changed with the UW System (UWS) vis-a-vis the State which resulted in the pursuit of a unified plan. For two years, a process has been underway that has included a functional request, technical option identification, technical specifications and a solicitation to three major vendors.

The vendor proposals were gauged on reliability, responsibility and cost. Negotiations are currently underway with WBAA for a cost-effective, scalable, IP-based network. This process was quite different from the regular RFP process. The steps were iterative and grew increasingly specific as the process unfolded. It was also possible to dialogue privately with each vendor, which isn't the case with a standard RFP.

Converged Network Services

Converged network services have been the next thing for a long time. Now they are here, with the exception of voice which is a number of years out because of large scale economics. Voice over IP (VoIP) will probably serve small office applications in the meantime. The converged services include:

  • Video networking
  • Wide area networking (WAN)
  • Internet transport services (without ISP services)

TEACH Subsidies

A major objective of the business case for the converged network was ubiquitous access for PK-12 institutions and libraries on a statewide basis despite their constrained budgets. Therefore:

  • Their bandwidths up to 1.544 Mbps ("data links") will be subsidized at $100/month
  • Their bandwidths greater than 1.544 Mbps ("video links") will be subsidized at $250/month

A link is synonymous with an access circuit, i.e., a service delivery point. The network has three layers:

  • End user interface (first layer)
  • Aggregation point (second layer with 17 statewide access nodes)
  • Core network hardware and the capacity necessary to deliver the application to the service delivery end point (third layer with MPLS and a core network switch in each LATA)
  • Service delivery end point (basically the same as the first layer)

BadgerNet Managed and User Managed Video

The common elements of both BadgerNet managed and user managed approaches to video networking are:

  • Increments of bandwidth from 256 up to Gigabyte Ethernet as desired by the institution. The goal is to implement bandwidth changes in a week or less.
  • Network infrastructure of aggregation nodes and core nodes that are sized by the vendor to be non-blocking.
  • Quality of service features available.
  • High priority and low latency applications possible (some state agencies have low latency data applications).
  • Virtual private networking available between pairs of agencies for data and/or video on a non-metered basis, e.g., a flat rate UW System intranet between campuses.

These features allow for the projection of costs on both an annual and biennial basis within a stable five-year window (not including any ISP costs). The differences between vendor-managed and user-managed service, in addition to cost, are summarized in the following table:

 

Vendor-managed

User-managed

Video

 

 

1x3 video session

Yes

No

h.264 codec

Yes

No

Scheduling system

Yes

No

MCU/Gateway

Yes

No

Continuous view

Yes

No

Continuous audio

Yes

No

Help Desk

Yes

No

Network and Codec maintenance

Yes

No

Free Site Controller Upgrade

Yes

No

WAN and Internet

 

 

Routers

Yes

No

Today, every user group, including the UWS, is saying "We'll tell you how much we'll buy when you tell us how much it will cost." Any Internet traffic will go to a core aggregation router, but it won't go across the core network because there are ISPs at each of the 17 core nodes. There are up to eight Gigabyte Ethernet ports per LATA, and WiscNet is guaranteed to have one of those ports in each LATA.

The network topology will be discussed in greater detail at a technical meeting next Monday that is open to participation from UW System institutions. The network solicitation only specified the location of the four core nodes to meet the needs of the UWS. It did not specify the aggregation points because it was up to the vendor to chose those such that it could deliver postalized rates across the state.

Back in the late 1970s, the philosophy was to let video networks develop organically around local, primarily geographic, needs. This has been fairly successful and this new network plans to enhance it. It will include sophisticated scheduling, including inter-network scheduling for schools in different consortia.

WAN Services

There are several possible WAN service scenarios:

  • User managed VPN intranet links can be set up between sets of institutions.
  • Internet traffic can be sent via VPNs to WiscNet and/or other ISPs
  • Video traffic can be routed via VPNs through a Multipoint Control Unit (MCU) from either a managed, unmanaged or external source. The MCUs are sized to handle the maximum amount of traffic to and from external entities. There will be four MCUs, one in each LATA. The level of need still needs to be quantified, e.g., the volume of university classes for K12 institutions.
  • Point-to-point user managed video links are also possible without going through the MCU.

The feasible number of VPNs in the network can grow to about a half a million. It is not yet clear if unused VPN capacity can be repurposed on the fly to handle bursting from other circuits. A lot of effort is going into writing Service Level Agreements (SLAs) for the VPNs

Contact Information:

Network Update (Ron Kraemer)

SBC is viewing the UW System as a big customer. DOA announced the intent to negotiate with the WBAA in late October. On December 7, UWS President Reilly met with two SBC presidents. President Reilly told SBC that faculty, staff and students need to be kept happy. SBC was given usage data and they were surprised by the magnitude. A follow up memo from President Reilly asked for a technical point of contact that CIOs could work with.

Many issues remain to be worked out, which is probably why the contract signing has been delayed until 2005. Nevertheless, a network engineering meeting will be held on December 20th. About 25 people from UWS will attend along with another 35 or so from school districts, CESAs, Technical Colleges, etc.

DOA has requested that each UWS institution identify a "conversion point of contact" by December 30, 2004. DOA plans to provide that person with a form to use to respond by January 14, 2005 to questions that were listed in the BadgerNet Newsletter. These forms will include commitments for at least the initial bandwidth purchase. It may be more advantageous to respond as a group, rather than as individual campuses. Brian Remer will send DOA the names of the CIOs as their points of contact. When more information is available, a CIO Council teleconference will be held. Meanwhile, the WiscNet Board has set new bandwidth rates. The cost per unit of bandwidth has been steadily declining over the last five years.

Learn@UW Utility Update (Kathy Pletcher and David Wirth)

There D2L upgrade to 7.3 release was difficult. There are additional features, but some performance issues remain with large class sizes. The 7.4 release is supposed to have optimized gradebook code that will improve performance for classes of 300 students or more. D2L has not yet issued a release date. The 8.0 release will have a .NET code base.

The UWS institutions would like greater availability for reporting, and greater facility for moving data from development to test to production. Without an export tool, it is difficult to move a course from one place to another because movement requires D2L involvement. An export tool is also necessary for archiving courses. When courses were initially converted they were all active by default and there was no batch process to de-activate them.

The various subcommittees are still meeting. The Training subcommittee is focusing more on best practices rather than simple literacy. The SLA/Hosting subcommittee is working on a new agreement with the input of Learning Innovations and UW-Milwaukee (UW-Milwaukee will be moving to Learn@UW in fall 2005). There will be an annual SLA review process that culminates in time for the fall semester. The Functional Request subcommittee has a new list of items that need to prioritized.

Learn@UW has been focusing on patches and upgrades. On January 6, 2005, a patch to improve the gradebook will be installed.

The Implementation team is developing policies for archiving gradebooks, discussions, etc.

An end user survey is being developed for the spring based upon the one in use at UW-Milwaukee. Questions will be given to the campuses to fold into whatever surveying practices they already have.

The Breeze Live pilot is running on a server at UW-Stout. There has been little feedback either positive or negative.

This afternoon, the site administators and LTDCs will have a conference call to discuss campus issues, which largely have to do with communication. Learn@UW has created a D2L course site that will be used to post lots of information, e.g., SLAs. Interested people will have to request accounts from learnuw-utility@lists.services.wisc.edu. There is also a weekly email from Learn@UW with important announcements, news, etc. Anyone can subscribe to it by sending a blank email to subscribe-learnuw-announce@lists.services.wisc.edu.

With all but one UWS institution reporting, current usage estimates include 6,762 courses and 181,698 students on both the Learn@UW and UW-Milwaukee servers. Several institutions are creating shells for every course whether or not they actively use D2L. If sites are created for each course, it may be best to let faculty members activate them as they desire.

Grade Integration

The D2L to SIS grade submission pilot is underway at UW-Green Bay and UW-Madison. Endusers and functional officers at both institutions brainstormed and developed the functionality via a highly collaborative process. There was also good collaboration between the MILER team and Learn@UW. The experience at UW-Green Bay has been positive. They spent about 230 hours of local development work and testing with an eye to benefiting the other UWS institutions. Subsequent rollouts will likely require less than 20 hours of local campus work. All faculty who wish can now enter grades via D2L and click two buttons to approve the transfer to the PeopleSoft grade roster. It is important that faculty use the same grade model as their PeopleSoft implementation, e.g., ABCDF. This semester's experiences will be used to inform the planning for the rollouts across the UW System.

ORACLE/PeopleSoft update (David Hart and Tom Scott)

As of two hours ago, ORACLE owns 51% of PeopleSoft. The ORACLE/PeopleSoft situation has been evolving for 20 months. A Collaterals working group has been continually tracking the developments. Their conclusion is that it is important to position the ten UW System PeopleSoft campuses to move quickly and carefully to version 8.9, which is the most supportable base for going forward. Until now, most UWS institutions have been looking toward implementing in summer 2006 so they could benefit from UW-Madison's experience in summer 2005. Recent developments may increase the urgency.

ORACLE and PeopleSoft have significantly different ERP architectures. PeopleSoft ERP systems are database agnostic, but obviously not ORACLE ERP systems. A new ORACLE application architecture will be developed over the next three to five years. Meanwhile, ORACLE has committed to carrying the existing architectures forward. The PeopleSoft architecture will be "supported" for ten years. ORACLE has also committed to releases of PeopleSoft Tools X next summer and PeopleSoft Scout, version 9, in December 2006. Regulatory and legislative functionality will be included, but it is not clear what additional features or enhancements will be developed. Presumably, the ORACLE ERP customer base will be pushed toward the new architecture as soon as possible. It is not yet clear whether moving from PeopleSoft ERP systems to the new architecture will be an upgrade, or a complete reimplementation.

MILER is built around a PeopleSoft technology stack, and UWS institutions will probably be given incentives to move toward ORACLE. Some may wish to go that direction. Meanwhile, some UWS institutions may be happy to continue running PeopleSoft indefinitely. To continue running PeopleSoft ERP systems into the future, support for PeopleTools needs to be available. The UWS has it licensed in perpetuity. Perhaps companies of PeopleSoft engineers will spring up from consulting firms.

PeopleSoft Grants and Projects

UW-Madison, Milwaukee, and Extention will be acquiring the PeopleSoft Grants and Projects module, which will result in beneficial restructuring of PeopleSoft maintenance costs for all UW System institutions for all PeopleSoft products for a few years. Once the contract is signed, details will be distributed. Lorie Voss and Tom Scott did a tremendous job.

NCES proposal to collect student unit record data through IPEDS collection process (David Hart, Gail Bergman and Sue Michalek)

If it should become necessary to move from the current CDR to a unit record type of reporting for students, it will require a significant effort. Presumably, the current analytical system would be combined with a transactional system. If Congress mandates the NCES proposal, it would still need to be accepted and approved by the Office of Management and Budget (OMB). Currently MILER provides a PeopleSoft extract that feeds into the current CDR. The new request would likely include requirements for additional transactional and end-of-semester data that is not typically collected at present. The proposed fast-track timeline includes a mandatory pre-implementation pilot study for certain institutions after OMB review in 2006-07. The functional and operational offices at each UWS institution are being kept in the loop by OPAR.

IPEDS data

The 2003 IPEDS employee data that resulted in increased ORACLE license costs had "quite extensive" erroneous counts due to the instructions given to the programming staff and the programming itself. The programs are being re-written. Fall 2004 employee data is being collected, at which time the fall 2003 data can be revised. Any changes in the overall numbers, up or down, will reflect correction of errors in the summation of original data that were generated at individual campuses at different times.

Report from the Common Systems Review Group meeting (Meachen, Kraemer, Berens & Stunden)

The APBS project sponsorship has moved to Darrell Bazzell at UW-Madison and the steering committee will now include CIOs David Dumke of UW-Stevens Point and Bruce Maas of UW-Milwaukee.

The short term goal is to create a project plan by the end of January that will include expectations, timelines and budget targets. Ron Kraemer and Diane Haubner will participate in that effort. The Common Systems Review Group (CSRG) has given approval to continue spending money at the current rate until the project plan is delivered and the project can be reassessed in light of the goal of going live on April 1, 2005. Meanwhile, the system test is yielding a fair number of level 1 and level 2 errors.

Cancellation of IT Summit; scheduling of CSRG Retreat (Ed Meachen)

Given the APBS developments, having an IT Summit in January was deemed to be ill timed. Instead, on January 13 and 14, 2005 there will be a Common System Review Group retreat to look at the issues raised in consultant Don Norris's study. The goal is to develop a shared vision, process and structure that will be brought to a possible springtime summit of CBOs, CIOs, provosts and possibly chief student affairs officers for structured discussion and vetting. Subsequently, the plan will be brought to President Reilly and the chancellors.

Moreover License (David Dumke)

Moreover Technologies has made an offer of under $25,000 with a five year commitment on increases. Their product is an RSS news feed that is often used in portals to deliver 300 news sources in a multitude of categories. David Hart will coordinate the license acquisition before the end of the year. Moreover could possibly be run at FASTAR.

Report from CBO Retreat (Ed Meachen)

A major effort at the CBO retreat focused on generating examples for the Legislative Audit Bureau of efficiencies gained through not filling various administrative positions.

The Croquet Project (Dr. Julian Lombardi)

http://www.opencroquet.org/index.html

Julian Lombardi runs the UW-Madison DoIT Learning Solutions group. Much of what we do with computers is not much different than looking at pieces of paper. We're not truly interacting in a virtual world. Dr. Lombardi's work is being done in collaboration with a number of fascinating people, including Alan Kay and the developer of gopher. The initial question was, "What would we do differently if we could go back 20 years and redo everything knowing what we know today?" Back then there were computers of limited capability without ubiquitous network connectivity or multimedia experiences. Today, we're not taking advantage of the full range of available capabilities, and are proliferating the document metaphor on our computers.

Why did we travel here today to meet in person, instead of somehow having a meaningful experience online? Perhaps because we can't do that with a page metaphor and our noses pressed into the desk. Pages are great, they belong in context, just as they are in the context of our social interaction in this real world meeting.

The campus is not about the information resources that we impart to our students, i.e., not about our "learning objects". The university is all about context. That's why we put pictures on the opening pages of our university websites. MIT gives away all of its online course content because, ultimately, the content is valueless compared to the value of the faculty and the community at MIT. Focusing on content misses the true value proposition.

People do not have access to higher education unless they give up their lives and go to the physical institution. Online institutions, as they are constructed today, still miss the essence.

The first step is building context. Dr. Lombardi demonstrated that, using Croquet, a Mac desktop can be virtualized on a PC and put in a massively multiuser collaborative environment in real time - and vice versa. I.e., each user can see each other user's "desktop" at the same time within a broader context. Croquet runs on Linux, Macs, PCs and ultimately handheld devices. The virtual machine can be burned onto a chip to obviate the need for an operating system.

Multiple Croquet users, and their computers, are connected via the network technology T-Time, which does not require a server. It is similar to the gaming technology that students are familiar with. Note that today, more and more faculty want to "mod" shoot-em-up games to use them to teach, which tends to be an isolated one-off effort.

About twenty institutions, and people, are working on Croquet around the globe, including Alan Kay, David Smith (Tom Clancy games), David Reed (HP), Andreas Robb (suite programmer) and Mark McCahill (gopher).

When you are in a Croquet space, your identity is known via LDAP. A camera on your computer would allow you to be your own avatar in the space. Either a professor or a student could manipulate the space simultaneously via serverless communication between the machines. A unique capability is the ability to create (author) a new 3D space that is immediately shared between the two (or more) people. Entering the new 3D space is like stepping (hyperlinking) into a looking glass (portal). The ability to move between virtual worlds is not seen in any other technology today. When you finish with a scene in a game, you're presented with the next space. Croquet is, in effect, duplicating the framework that gave us the web, but doing it in 3D for advanced simulations and rich media without shoehorning plug-ins into browsers. You could step from the Civil War to Neptune to Iraq.

A pre-croquet, single user chess game can be projected into a collaborative space and played (in turns) and observed by multiple people. Multiple people can similarly drive the same web browser, which allows for shared problem solving. Legacy applications can be exposed to multiple people with realtime VoIP enabling a communications overlay, e.g., a help desk service. Flash animations, web pages, movies, text, etc., can be painted onto polygons within the 3D space in real time and controlled by any of the participants at any time.

3D "objects" that are created in Croquet are automatically pushed up to a server where they are available for everyone in Croquet-space. Objects are also instantly published in a peer-to-peer sense to the other "nearby" collaborators.

Gates can be built to prevent the unauthorized from entering spaces and to change the attributes of the user as they enter the space, e.g., they change into a fish when in an aquarium space. This allows for copyrighted media to exist safely within a space.

Anything that can be done in a modern computer game can be done in Croquet and modified during runtime by any of the collaborators.

When within a space, a model of that space can be made automatically and carried out through a portal into another space/world. I.e., a faculty member could look at live models of all of their learning environments and see who was in each in real time. If they chose, they could then step into those worlds and join their students.

When you initially launch Croquet, it goes to the worldbase server to find out where the interactivity servers are that tell you where you are in your local area. In effect, the world builds out from where you are. Your Croquet queries the interactivity server to tell you who is in the space with you. You can place a new persistent object in the immediate locale, which gets promoted to the interactivity server for the benefit of nearby collaborators, and then to the worldbase server, which is just another peer to peer level on really fast pipes. At this point, every Croquet user has a copy of the object available. Collaborative, whiteboard images can be extruded, deformed, etc to develop objects inside Croquet.

>Instructional designers at UW-Madison have taken D2L content and put it in Croquet and created a "course" (literally) of study for learning material while simultaneously moving through space. It is where museum theory meets instructional design theory: a constructivist learning environment instead of a post-industrial model.

Croquet is open source and can be downloaded, however it is still in a development state suitable for "raw-meat-eating programmers", not the faint of heart. User interfaces are being built to make the environment less intimidating. Hopefully, all of these tools will be available for a beta release in summer 2005.

Next meeting

The next CIO Council meeting is scheduled for Thursday, February 17, 2005.