Project to evaluate a video-enhanced collaboration product (original) (raw)
This page defines a Kan-ed project to evaluate the feasability and potential value of providing a "video-enhanced collaboration" service within Kan-ed.
Project scope:
Conduct tests, and encourage testing by other groups, of the Marratech video conferencing system in order to render a recommendation for deploying a video-enhanced collaboration service within Kan-ed, and assess the suitability of the Marratech collaboration product for such services.
Note that this project does NOT extend to deployment of a productional service, but only to the assessment of the desirability of such a service.
Business case:
Article 72 of Kansas Statute 75-7224 instructs the Board of Regents to "assess the need of schools, libraries and hospitals for full-motion video connectivity. Based on its findings, the board may develop a plan to provide such connectivity."
The full range of "full-motion video connectivity" may be broken down into a collection of video services, including (though not necessarily limited to):
- Video conferencing services with no or little simultaneous data exchange (as with whiteboard, chat, or applications data),
- Video-enhanced collaboration services, also known as "rich-presence" or "rich-media" services,
- One-to-many broadcast services with or without simultaneous data presentation,
- Streaming video-on-demand services also with or without simultaneous data presentation.
These services can be provided using hardware customized for video processing (sometimes called "codecs") or by using software running on desktop PCs, and combinations of such equipment. For groups of around 5 or more participants, some hardware and/or software that relays traffic among participants (such as the Multipoint Control Unit in the H.323 model) is usually required. Such relays are typically implemented as either customized hardware or as software on commodity PCs.
Room-based systems such as PolyCom video conferencing units are examples of the first category above. The Marratech client-server products fall into the second and third categories above, when they are used to provide a "rich-presence" or "rich-media" video conferencing platform for both small groups of 2 to 20 participants, as well as groups of several hundred clients, connected via unicast and an unlimited number of clients connected via multicast.
The Marratech product suite includes cross-platform desktop clients (PC, Mac, and Linux) which communicate with one another through a server running (cross-platform) software that relays audio/visual, chat, whiteboard and applications sharing traffic.
Given appropriate management facilities, such as record keeping, and an appropriate charging model this product could be offered as a (partially) self-sustaining revenue-returning Kan-ed service. Alternatively, it could be offered to add value to Kan-ed network connectivity, and to demonstrate the superiority of Kan-ed over commodity Internet connectivity, since it is anticipated that the Marratech products will operate much better over the high-bandwidth Kan-ed backbone.
Roles:
Plan sponsor:
Rightmeier will take recommendations to the Management Group.
Management Group:
Gardner, Rightmeier, Williams, and Robinson will decide whether to accept the recommendation, ask for more information and/or additional study, and decide whether to deploy a rich-presence service (and with what configurations).
Stakeholders:
The Kan-ed user community, and especially the User Advisory Committee, who will advertise the test to their constituents, collect anecdotal feedback from testing conducted by their constituents and themselves, and provide advice on the test approach and deployment of a service.
Project manager:
Grobe will manage overall project activity from planning to assuring products are delivered as scheduled.
Plan summary:
This section outlines the overall plan and identifies non-personnel costs along with personnel effort required to produce end products.
- Acquire the Marratech server software license for 3 to 6 months
Marratech is a suitable platform for testing such a service for a variety of reasons, but particularly because it is the only Internet-based cross-platform (Windows, Apple, and Linux) e-meeting product to combine high quality voice, video, whiteboard and application sharing for more than 5 participants over secure, encrypted connections
License costs are around $2350 per month, payable quarterly in advance, and it is possible to roll these monthly payments over into a permanent license (which breaks even with the monthly purchase rate at around 18 months), should the Management Group elect to do so. - Install the server on Kan-ed hardware
Discussions with Kan-ed IT staff suggest the hardware platform hosting the current Kan-ed Listserv server would also provide an adequate platform for hosting the Marratech server during this test.
Running a conference with 15 simultaneous participants could nominally generate as around 10 Mbps of inbound traffic. Running an auditorium will generate one outbound stream for each unicast connection, so this could easily present a problem depending on the new circuits to connect KBoR to Kan-ed.
Funk and Grobe would do the product install (Estimated effort: 4 hours.)
Grobe would administer the Marratech application though its built-in Web interface. (Estimated effort: 30 hours.) - Establish a web-presence and Listserv group devoted to this test
Funk would define the Listserv group (Estimated effort: 1 hour), and Grobe would build the Web presence (Estimated effort: 15 hours over the duration of the project.) - Announce the test via:
- the test project web site,
- personal communications
- the UAC,
- etc.
Among other information Kan-ed staff MUST make users aware that very few of them will actually reach the service via the private Kan-ed network, which is to say, product performance will almost certainly be impaired.
(Estimated effort: 4 hours on the part of various Kan-ed staff.)
- Conduct user training and testing sessions or arrange for such sessions to be conducted by contract groups and/or interested parties.
This would necessarily include training for the NOC Tier 1 and/or Tier 2 support group so that they would be aware of the project and possibly assist users with problems.
(Estimated effort: 5 hours plus travel and organizing time for each such testing session per Kan-ed participant. 3 hours for each NOC representative attending training. An average of 5 hours over the duration of the test for each NOC member providing end-user assistance. ) - Collect user feedback on client and server usability, client and server performance, and implications of any limitations identified in system management functionality.
Some specific items of interest with respect to the client:- Reliability
- Ease of installation
- Completeness of the cross-platform implementation
- Ability to support cameras, mics and headphones
- Quality of video
- Repeat previous 3 items for audio, chat, and security functionality
- Adequacy of video for collaboration purposes
- Ease of use of video features
- Overall ease of use
- Adequacy of online documentation
- Adequacy of vendor support and training
Some specific items of interest with respect to the server: - Adequacy of capacity (number of users, bandwidth, etc.)
- Adequacy of scheduling features
- Ability to dial out to telephone-connected clients and telephone bridges
- Ease of installation
- Economical use of hardware resources (CPU, NICs, memory, etc.)
- Adequacy of online documentation
- Adequacy of vendor support and training
Some measures of usefulness of the service - General per person cost of using the client, including training, peripheral equipment, etc. per conference attended
- Level of anticipated savings due to use per conference
- General usability of the service
- Desireability of such a service
- Breakeven price point
- Suggested charging model
(Estimated effort: Grobe, 20 hours.)
- Prepare a detailed description of server facilities and operation, addressing management of conference room access and security, session recording facilities, session scheduling, etc.
(Estimated effort: Grobe, 10 hours.) - Prepare a recommendation to the Project Sponsor and Management Group concerning deployment of a rich-presence service and the suitability of the Marratech products as the underlying platform. This would require
(Estimated effort: Grobe, 20 hours.)
Start and Stop times
This project will begin as soon as possible with the major impediment to starting being acquisition of the Marratech server license.
A recommendation will be produced within 6 months of product acquisition.
Assumptions
- The identified server has adequate capacity for the load experienced by the Marratech server
- The Listserv service running on the same server will not be degraded
- Network bandwidth to the Marratech server will be adequate for the load
- Network bandwidth available to other LAN activities will not be significantly degraded
- The Kan-ed Groupware server provides an adequate platform for hosting the Marratech test web site (http://www.ksbor.org/\~mgrobe/marratech).
- NOC Tier 1 and 2 personnel would be available to assist in roles identified above.
- Users will be found to test the client on all three desktop platforms.
- We can resolve usage conflicts arising from the lack of a built-in scheduling facility.
- If a decision is made to deploy this service, the server will probably be reinstalled on a new server platform located in a different venue, probably the Kan-ed NOC, and MAY cause a difference in access procedure affecting end-users.
Risks
- If the local Kan-ed server has insufficient capacity to handle both Marratech and Listserv functions, we would probably have to acquire another server as soon as possible, thus introducing a delay in the project time line.
- If the Kan-ed server has insufficient bandwidth for connected users, or the service interferes with other applications sharing available bandwidth we would probably have to acquire another server and negotiate a contract to install the new server in the Kan-ed NOC.
- User demand might overwhelm the test configuration, leading to lower user satisfaction.
- Since few users would actually be traversing the Kan-ed backbone to reach this service, their experiences with the service will be confused with the performance of their commodity Internet service rather than the Kan-ed network and the product itself. This MUST be made clear to users.
- Changing location of the productional server MAY pose a hardship for the user community.
Please address inquiries about this document to its author:
Michael Grobe
Kan-ed
785.817.2992
| 
1000 SW Jackson, Suite 510 Topeka, Kansas 66612 785-296-0843 phone 785-296-7052 fax | |
| --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | |