The great email outsourcing dilemma

One of the current memes in university computing is that of outsourcing email, or more accurately outsourcing your communications and calendaring platform.

There are various variations on the proposition but this is one version:

We need an integrated email and calendaring solution for all members of the university to both schedule meetings within the professoriat efficiently, and to communicate timetables and tutorial schedules to students, including late changes. We also need a viable and robust way of communicating by email, and an instant messaging solution would be nice for presence information. Oh and it has to be web accessible.

Well, you could do it all with Microsoft Exchange. That suggestion lasts just about as long as it takes for someone to work out the licensing costs. Usually about 10 minutes. There is also the problem that exchange only really supports Microsoft email clients running on Microsoft desktop operating systems (Entourage aside) so we need to invest in web infrastructure to ensure ease of access for non users of Microsoft operating systems.

Entourage would provide a solution for OS X users assuming that the have Microsoft Office, and have bought into the whole Microsoft ecology. Which given that they have Macs, may not be the case.

So, given that we're not email virgins and already have a solution someone will then suggest continuing to run an in house alternative that gives the same functionality. There's only really two possibilities - Zimbra or Sun Java Communications Express.

Both do the same things, provide a nice web client with integrated calendaring and instant messaging, and becasuse they are basically imap servers allow you to carry on using the desktop thick clients you know and love - yes you can still use pine. More seriously there are connectors for outlook to allow syncing and by extension syncing to your mobile phone.

The iGeneration doesn't care about any of this as they think of email as a web application , and quite often have sexy phones with built in browsers.

Doing it in house means you need to provide infrastructure, disk space, mail system administrators and the rest. It also looks like a contained system, ie one you can draw boundaries around which makes it easy to both identify the costs and to outsource.

And when you look at Zimbra or Sun JCS, what they do looks a lot like Gmail Plus Google Calendar, or indeed Windows Live! (aka Hotmail).

Interestingly Yahoo! who now own Zimbra offer a hosted version for a fee - ie you use Zimbra running on Yahoo servers rather than your own for a fixed cost. This fixed cost at least makes the costs predictable and constrained. Equally interestingly almost no university has gone down this route.

Universities who have outsourced have outsourced to either Google or Microsoft, and without being snarky, most universities you might have heard of have gone Google. The major exception that comes to mind is the University of Queensland which has gone Microsoft.

So, what do Microsoft offer?

The honest answer is I don't know. What I can say is what they were offering a few months ago was interesting but not persuasive. Essentially it was Windows Live! logo'd with your University id and the promise of future integration with Exchange . The model being that you give the students the webmail client, staff Exchange, and that this integration layer allows tutors to put sync tutorial calendars with individual student calendars. It also gets rid of the task of having to provsion thousands of new undergradauate accounts at the start of the year and delete them once the individuals concerned have graduated.

Sounds good, except that the student calendars might be better provided via the learning management sytems, and that universities are not hotbeds of pro microsoft orthodoxy. However the model gets round one of the real problems with outsourcing email - which is where is it stored.

This is important as much of the business of a university is conducted by email and there is a need to retain this information within the jurisdiction required, plus arguments about the intellectual property contained in emails. If the staff email is kept inhouse these problems go away. We probably don't care about student email, all we care is that they have an account somewhere and read it regularly as it provides a means of communication. Providing an email account gets round the problem of ensuring that they have their own. We could just as well require them to get an account from one of the free providers.

It also assumes that there is a clear dichotomy between staff and students. In a teaching based institution this is probably the case. There are a large number of students, relatively few staff and not that many postgraduates. Most students arrive, do their four years, and go, and if they're promising go to graduate school somewhere else. The overhead of providing exchange accounts for post graduates is probably not that great in this scenario.

In a research based institution it's different. There are a lot more staff, be they lecturers, researchers or whatever, and a lot more postgraduates. There are also a lot fewer undergraduates as a proportion of the community and it's a reasonable guess that we'd see quite a few moving on to do graduate school /research on the basis of a promising honours degree. Ie there's a continuum, not a dichotomy - and this is the reason that Oxford , for one decided not to outsource.

So what about Google?

Again they offer a solution that works for outsourcing student email with easy web access and a calendar. It's intersting that most universities that have outsourced email have basically only outsourced student email to Google keeping staff email inhouse. Again this is a model that works in a teaching institution, and has the advantage of not assuming (or precluding) the use of proprietary technology for the staff email solution. Given the antipathy to Microsoft in some quarters this is important and probably contributes to the acceptance of the solution.

And GoogleApps?

For third world universities, having Google Apps in the bundle would be a plus given that students may not own their own computer and may instead be accessing resources from internet cafes etc, while the university can not afford to provide enough in the way of public access labs. Being able to ensure access to a solution that will work witholder devices, or lowcost computing devices, but not having to invest in infrastructure would be a definite plus.

Non third world universities probably are working in a environment where students have a computing device of their own which supports a text processing application and a spreadsheet application - open office is more than acceptable these days.

Collaboration, as I have written elsewhere, is a chimera . Collaboration and sharing can be enabled in a number of different ways. If people require shared editing they will find and use a suitable tool outside of any formal availability - for example the zoho suite can be wrapped up inside of facebook - just as they will for a blogging platform etc etc.

And everything I've just said about Google Apps would equally apply to the Microsoft offerings in the Windows Live! or any other bundled applications offering.

So, if you have a distinct student population with relatively little transition to staff and graduate status you can outsource email with relative ease. Its a constrained problem. It gives students a web based experience similar to that which they are used to and removes the costs of provisioning these accounts and may produce a cost saving. Due to students being handled as if they were only consumers of resource we are assuming little continuation after their four years undergraduate study, and that as email accounts are only provided to enable a guaranteed communication channel we can assume that we do not care about any intellectual capital stored in their email messages. Outsourced email is basically an enhanced version of asking them to provide their own account.

Under such a model staff email can remain in house. These costs will continue to be required to be met by the university, but assuming the use of non propiretarry technologies, or ones without onerous licensing conditions te costs should be less and the disk costs should also be less - smaller is cheaper.

The moment we start to see a high degree of transition between undergraduate, graduate and staff status, and a high proprotion of individuals who contribute to the intellectual capital of the organisation the model begins to break down. Of course one could be radical and imagine that if the university was truly a community of free scholars one need provide nothing and they find it from their own resources.

Atractive as it is intellectually, in practical terms this is a step too far as universities need to hold and provide records, which predicates the provision of central services to some extent, and the need to conserve intellectual capital reinforces that.