UC Berkeley currently receives the bulk of its voice service from AT&T. The product that we use, Centrex, is phone service delivered from the AT&T central office. In many cases, getting service from an external provider may make sense. In the case of the campus phone system, we have sufficient scale to operate our own service and estimate that we will be able to save somewhere between 15 and 20 percent over current costs by doing so. To provide our own phone service, we will be installing equipment on campus that is typically referred to as a Private Branch Exchange, or PBX.
While UC Berkeley has enough scale to reduce operating costs for voice, we feel that there is even more savings to be gained by combining our operation with other UC entities (Office of the President, campuses, and labs). The first step will be to combine the operations of the phone systems of UC Berkeley and Lawrence Berkeley National Laboratory (LBNL). Other UC entities are likely to find the cost model associated with a large combined operation attractive, and we expect that we will gain further savings in the future.
A limited deployment of the PBX is underway. LBNL will begin actively assisting with managing the voice infrastructure starting July 1, 2009. An aggressive deployment of the PBX is planned for October 2009.
Berkeley's new voicemail system will work with both the PBX and the Centrex services so there will be no need to change voicemail systems as we deploy the PBX.
Voice services and innovation
There are a number of innovative services being offered in the marketplace. Many are focused on integrated messaging or using existing personal computers to handle voice communication. Some individuals or departments may find that one of these services provides the right mix of cost, quality, and features. Among the challenges that a university faces in deploying services on the bleeding edge of technology are: expense, quality, and life-safety. Many innovative communications services are based on sufficient, but not reliable or high-quality, service that may not be adequate for all purposes. In many cases, new services explicitly advise folks not to use them in emergency situations where there is an expectation that a 911 call will provide an accurate location of the person calling for help. In most cases, the costs for software and hardware that support advanced Internet-based services are considerable but can be recovered across a very large installed base or through advertising revenue.
Even with all of the services that are available in the marketplace, there is still a place for basic phone service. Quality, reliability, and integration with emergency services are some of the reasons that not everyone has abandoned a wired phone line. Given that there are plenty of options for mobile and Internet-based phone service with advanced features, it seems to make the most sense for the campus phone service to be focused on basic services at the lowest possible cost. Individuals and departments wanting more advanced service can obtain those services from external providers. Over the next few years, the main objective for campus phone service is not innovation but cost control. The campus community should expect to see a very basic service and a couple of options for more complicated applications, such as call centers or multi-button digital sets with call display, but limited features.
Voice over IP (VOIP)
In many cases, organizations with the right kind of infrastructure or telecommunications needs are finding VOIP a good solution. The campus phone system uses VOIP to communicate between modules and to other systems such as voicemail. Voice over IP to the desktop, however, will not be a service offered by the campus phone system for the foreseeable future.
While VOIP has many advantages, there are also some considerable limitations, mostly related to expectations of phone system operation in emergency situations. The main concerns with VOIP are that we expect phones to work when the power fails, and for emergency responders to find us when we dial 911.
Traditional phone systems are powered from one or a few locations and backup power and emergency generators are focused at those locations. Data networks, on the other hand, are very distributed and the majority of devices that attach to them are locally powered. In order to ensure that data connections and VOIP phones remain active during a power failure, UC Berkeley would need to make a considerable investment in battery backed-up power systems across campus and install equipment to power VOIP phones during power failures.
We also have a responsibility to our community to provide a reasonable level of life-safety systems. While one of the big advantages of VOIP systems is that devices can readily be moved around, this advantage can present a problem for emergency services. Without investing considerable time and money into a phone location registration system and related software to enforce location registration, we would quickly lose track of the location of phone sets and emergency responders would show up at incorrect locations.
Key and departmental systems
In the past, because we have focused on providing a wide range of services, it has made economic sense for departments to install local systems to save money. Supporting these distributed systems is expensive. In the future, these systems will not be supported and will be prohibited from connecting to the campus phone system. The cost of support is only one consideration; the other is safety. Phone systems need to accurately report the location of persons dialing 911. Many systems designed to support small groups or departments don't have this capability. By providing inexpensive phone service, we hope to reduce or eliminate the financial incentive for groups to purchase their own communications equipment. Where local equipment is necessary, arrangements must be made with external vendors for support and connection to the public phone network.
More information
If you would like to know about this project, send email to