We used to think of client computing as a desktop computer with a software build on it. That's changing, and the old view is no longer sufficient. Trends changing it include Windows 7, Virtualisation, Macs, App Delivery, Cloud services, Smartphones, Tablets & Consumerisation.
Three different ways in which you can use virtualisation like techniques for client computing:
- Terminal Server (fairly familiar).
- VDI (Virtual Desktop Infrastructure). An alternative to terminal server, as your desktop has complete control over the VDI Virtual machine. In theory less contention between users than with terminal server. I got the impression James thought VDI technology too new for production use,, more something to keep an eye on.
- Client virtualisation - VM on client machines. eg run Windows on a Mac, or VM for personal use on a work laptop. This lets users install things they want under their own control, but without messing up the locked down corporate image.
Hardware Thin Clients.
Is the promise of hardware thin clients too good to be true? Is the cost of ownership *actually* lower? No, not when you include the cost of the data centre infrastructure. Including that, the cost savings if any are marginal. [This is particularly true at Bristol, where we are struggling for data centre space]. Where are hardware thin clients most sensible? In a new building, where you can design the building with less requirement for aircon due to the lower heat output. Where are they least sensible? For student labs - students are intensive users, and thin clients are unacceptable for students if they can't watch Youtube!
Macs are unquestionably becoming more popular. [Figures of 1/4 to a 1/3 Mac ownership by students quoted by various institutions]. Macs are gorgeous! There is nothing you can do to stop people buying them. Luckily, some new technologies make it easier for us to support them. There are options to deliver Windows apps to Macs. And if your data is in the cloud it is as easy to access on a Mac as on a PC.
OSes don't matter to people, applications do. Apps are what people use. Smartphones have changed people's expectations of apps. Apps from the App Store arrive on your phone within a few minutes. In corporate settings, it takes forever to get an app approved, tested, and deployed! Our application delivery mechanisms need to get quicker. Yes, this will be tough on us, but from the users point of view that is not the point. So how do you do it? Offer an App Store like experience on the desktop. Citrix already have one - the stupidly named Citrix Dazzle. Chose your app, potentially enter an internal charge code if it isn't already licensed, and then it's there. [Maybe within a day, if not within the 5 mins of the phone, but better than the months it can takes us at the moment].
MS AppV (formerly SoftGrid). VMWare ThinApp. Separates the OS from the app. Can allow installation of two versions of the app.
Cloud Based Services
eg Dropbox. How many people here use Dropbox? Lots of hands went up (don't tell your data protection people!). But you use it anyway as it is *really* easy and useful. Whatever we say to users they will continue to use Dropbox. Telling them not to is just ignoring the risk, when we need to manage it. So instead we need to make our corporate personal and shared space as easy to access, and that means putting it in the cloud.
Smartphones & Tablets
There were 4 or 5 iPads in the room, but smartphones were ubiquitous. iPads are still too expensive and there are real challenges at the moment using them in an enterprise environment - manageability, app purchase, support, and getting data on and off them. But these things will be fixed within the next three years, just as access to corporate calendar and email has been fixed on the iPhone [remember three years ago nobody had one of those either].
An ugly word, but the right word to describe what is happening to IT. Enterprise IT remains a good thing, but consumerisation can't be ignored. Central IT services need to rise to the challenge, not ignore the inevitable. At the extreme, some companies have experimented with saying "just buy your own computer and look after it yourself". Just give the staff a cash budget. This definitely isn't the right solution for all users, but it is for some.
The support dilemma
We need to support a wider range of OSes. We need to support personal devices. Personal devices have always been there, especially in HE, so take advantage of it! It's a challenge. New services don't mean users stop using the old ones! The old ones are still valuable. People use Twitter & Facebook & MSN but still use email. Students are still queuing to use desktop labs. People do everything at once [at times during my travels over the last three days I've been using the iPad, iPhone and netbook simultaneously!]
Implications for staffing
New technology such as virtualisation brings desktop and server infrastructure closer together. This challenges our traditional split between systems and support. Client & desktop should be seen as an area in its own right, with someone taking clear ownership of the whole thing. [Remember users don't care about the systems infrastructure, they care about the client and the apps on it. But traditionally maybe we defined ourselves by the backend systems infrastructure, thinking that is the 'real job' - and in the process neglected the true real job, what is important to the users?]
A group organised around the client, responsible for delivering a common desktop. [Big caveat, this is still a work in progress, and they haven't yet delivered it!] They have a decentralised IT staff, but [will] have a centralised app packaging process available to these staff. Gold, silver & bronze levels of the common desktop.
Areas that need further work
- Backup/data integrity
- Remote wipe of mobile devices
- Print consolidation
- Green IT (- de rigueur to add this to list, but is it really important? He's a skeptic.)
The future is:
- consumerisation and cloud services - embrace these trends
- responsiveness, an app store approach to the desktop
- recognise that one size fits all doesn't work, especially in HE
- tablets and mobile devices will be pervasive, but thick build PCs aren't dead