An overview of the project
Instituted in 1987, the TARDIS Project is a student-operated computing facility, hosted by the Informatics Department at the University of Edinburgh. Our systems currently reside in the James Clerk Maxwell Building, which the Informatics Department intends to vacate in the near future.
The project provides free web hosting, e-mail accounts, mailing lists and other services to registered students and staff from the University. Thanks to our pool of local student system administrators, we are able to continually maintain and tailor our services to meet the ongoing requirements of our users; including services for student clubs and societies.
While we place great importance on the uninterrupted provision of service to our users, this in itself is only one goal of the project. We also aim to gather and foster a local community of practically-trained system administrators from the University of Edinburgh.
Students interested in system administration are welcomed and can join the discussion and development of the project, and those with some existing experience - though perhaps only in small-scale system administration - are able to work with a real, live computing facility with a large user base.
To assist recruitment of new members and day-to-day maintenance, we would ideally like to move Tardis to a location near George Square. Our physical security is somewhat limited at the moment, and a location restricted to Informatics students would be perfect, though we realize that this is a non-trivial requirement, and are glad to discuss the alternative options.
- Tardis used to provide training/tutorial sessions to students, by students and external speakers, for free.
This is a big point. We did have regular attendees and figures showing how many people came along. It provides both experience to students for improving presentation skills, and demonstrates an enthusiasm for people to learn off their own backs.
- We can provide a safe (Xen) introduction to new sys admins
- We provide a realistic environment for system administration with large numbers of users and large volumes of data
- We provide a realistic environment for learning and experimenting with real-world network technologies
- Skills users and admins can gain are not generally taught in Informatics courses, since they're somewhat more practical/vocational than theoretical
- We provide useful services for students (though people might potentially think it's wrong if these are provided long-term?)
- We provide useful services for societies with sys-admin capable members
- We have hardware such as the Cisco kit, the Suns, the Intel Xeon box and the rack, which students will not have access to elsewhere.
- A /24
Tardis exists to:
- Gather and promote system administration talent in Informatics
- Help the effective use and management of computer technology amongst University societies
- Provide an environment for discussion regarding sysadmin technologies (established and new)
Probably would be worth pointing out:
- That people have blagged jobs through Tardis contacts or Timelords or something.
- Timelord types make good contacts in general (or at least, Steve likes to give us hardware).
- That people feel they've learned something through Tardis, and that it makes for good CV pimping.
- That we have a quite surprising number of active users and societies using us (quite potentially 3 figures)
- People have used our services and environment to do their projects/dissertations. (We got another query about this kinda thing last week)
Counter arguements against tardis which we would expect to be presented with. If possible, a sensible rubuke to each argument would help in defending tardis's case if required.
- People have linux systems at home these days, they don't need tardis
Administering home-user linux systems is incomparable to running a shared multi-user installation with a number of live internet-facing services. The practices and processes required to run such sites are entirely omitted for home user systems. Tardis also provides admins with a suitable user base to appreciate usage issues with services they deploy yet does not put real world pressures on students.
- Why can't tardis just use a mesh network/vpn/blah to provide the same service?
A mesh network does not represent any real world situation and suffers a number of drawbacks. Whilst such a deployment would be of interest in its own right, it does not cover the situation which tardis provides.
Project members have dabbled in such networking projects in the past, but have found they get easily neglected.
- People have suitable resources from the university, why do they need tardis?
The university runs all their own systems. The act of running the systems is kept transparent to the students who do not appreciate the amount of work which goes in to maintaining those systems or some of the complex challenges faced in this process.
What do we really require?
- Xen hosts (how many?)
- What physical hardware currently exists to provide this
- Could an exchange be made, reducing space if additional hardware can be provided?
- Networking equipment
What would help us:
- A small amount of space (one rack) in an inf-student-accessible location
What we require, worst-case:
- A machine with network access, or funding for colo elsewhere
What we need to do:
- Get more people interested (ESSENTIAL!)
- Reduce our heat output and footprint