1. Research and Education is buying fiber on a phenomenal scale

2. Customer-owned networks offer the greatest opportunities for the future

3. Networks are becoming a critical part of the education experience

4. Public-private collaboration can overcome barriers to high-speed broadband infrastructure and create opportunities

5. The results of public-private collaboration are substantial enough to be disruptive

Introduction

The economic downturn beginning in 2000 has provided opportunities to build research and education networks. Since 2000 Research and Education organizations have been investing in thousands of miles of fiber, and have accumulated more than 21,000 miles to date.

This presentation will tell the story of Utah Education Network and our project to bring Ethernet and Optical based services into our network through a state-wide collaboration with municipal networks and a dozen incumbent and independent telecoms.

Background

UEN has developed and operated four different networks. One significant network is an analog/digital video network. Much of the backbone is analog, using a proprietary OC-48-like service provided by the local RBOC, Qwest Communications. The other primary service provided by UEN is a data network mostly comprised of T-1 circuits to in the first mile and a combination of OC-3 and DS-3 circuits in the backbone.

The video network is made up of nearly 400 sites with IP Digital video equipment and older analog based video equipment. The data network services all 10 state-run higher education institutions as well as several private institutions, 40 Public Education school districts, dozens of public libraries, and state government agencies.

The NGN 2000 conference marked a turning point for the UEN network. The message of the conference was that Ethernet would likely dominate through all layers of the network. Upon returning from Washington, DC, UEN immediately began talks with vendors. The initial response from the incumbent LEC was negative. Other carriers were contacted and seemed to be receptive to providing Ethernet services. These interactions were positive enough that UEN continued to plan and dream about conversion to an Ethernet backbone with Ethernet services in the first mile.

We realized that the ILEC was initially interested in selling existing services. The motivation to profit from sunken investments was strong. It was going to be challenging to get the attention of the ILEC.

Accomplishments

Ten months of planning and groundwork followed our initial Carrier inquiries. In the fall of 2001 UEN was ready to circulate an RFP soliciting Gigabit Ethernet service for the UEN backbone. We had planned a ring topology that would give us the ability to have 5 nines reliability while using a low cost Ethernet service.

It was a good thing that this project started in 2000. Had we attended the 2001 NGN Conference first, where GigE was hardly mentioned, we might have given up on ever seeing any results from our Ethernet efforts.

We received positive responses from vendors as a result of our 2001 Ethernet RFP. XO and Comcast were both awarded Ethernet Services contracts. Qwest and several other vendors weren[base ']t able to to respond because they didn[base ']t have GigE service offerings. The first GigE circuits were installed in the UEN backbone in Summer 2002.

Next we decided to tackle the problem of Ethernet in the first mile. The ILEC was not responsive to our ongoing request to explore Ethernet in this layer of the network. We moved our discussion to independent Rural Telecommunications providers in Utah. We felt confident that the independent telcos could be more aggressive because of their smaller networks and smaller, more agile organizations.

The first project was in eastern Utah with Uinta Basin Telecommunication Association (UBTA). UBTA had fiber in the ground very close to many of our education sites. They were eager to work with us and customized a new Ethernet service to meet our needs. This project consisted of 9 sites, each provided for $635 per month per site. After completion of planning and a request for bid, UEN contracted with UBTA for these services. The first Ethernet circuits were installed by UBTA in early Fall 2002.

UEN discussed these developments with our local ILEC sales team. The installation of a customized GigE service caught their attention. Coincidentally, a new sales team was assigned to service the UEN account.

This new team finally got the message of Ethernet services across to their upper management. First they pitched their transparent LAN service, an Ethernet-over-ATM-over-Sonet solution. The packets were not prioritized and UEN felt that this service would not meet our reliability requirements.

Finally, after months of conversations, the ILEC proposed a lambda-based solution. This turned out to be a better solution than a straight Ethernet play. Using Nortel Optera gear the carrier was willing to provide 8 optical channels running from our farthest north PoP to our farthest south location. UEN had contracted a decade earlier for video services provided over fiber in 105 of our sites, including all 10 higher education facilities. It was the presence of that fiber that created the opportunity for the new service.

An important quality of the Optera gear is that of being an optical extension cord. The only two requirements are that a service must be digital and it must be light. The ILEC devised a plan where UEN could move the proprietary video service onto on lambda and set up a second lambda as a greenfield GigE network. This migration was planned and installed in November and December 2003. At present UEN is planning the implementation of the greenfield GigE network for later this year.

Future Prospects

The UEN conversion to Ethernet Services and Optical capabilities happened much faster than any of us expected. Finding willing vendors took some time. Once these were identified we were able to leverage their participation with out other carriers. These projects have provided UEN with much needed bandwidth and has motivated our vendors to implement next generation services in their networks.

At the start we felt that our chances for success were not great. At best, we predicted that marginal progress could be made within four years and that it would take at least 7 years to make real progress. From our first inklings of Ethernet until full conversion of our backbone was less than three years. Our first access layer Ethernet circuits were installed in less than 2 years. We now have significant Ethernet projects under way in all areas of our state. Every School District, College and University talks about GigE and everyone is planning to bring Gigabit Ethernet services into their networks. It has been an amazing project.

GigE is making a real difference in the culture of networking in Utah. Districts are beginning to see the value of large amounts of bandwidth. Servers are being moved from the schools back to the District Offices. New projects are in the planning stages. Bandwidth is no longer holding back network upgrades.

We are also rethinking network convergence. Perhaps it will happen at a different level than the IP protocol. Time will tell, but Ethernet and Optics is allowing a whole different level of thinking to drive our network plans.

Conclusions

1. The NGN 2000 conference got us thinking about Ethernet services and planning for Ethernet in our network.

2. There is no question that the down economy in 2000 - 2003 really helped our cause. We had money when many other carrier customers didn't. Vendors, both of the carrier and equipment variety were willing to give us some attention.

3. Change happened in a much shorter timeline than we expected. We often laugh and quote the phrase, "be careful what you wish for, you might just get it. Then what do we do?" We find ourselves wondering what's next.

4. We have just begun to build our network. The implementation of Ethernet and Optics gives us many more options. Over time we will better appreciate the opportunities that await us.