opinion Right now chief information officers and IT managers right around Australia are facing a difficult decision regarding one of the most critical but also trouble-plagued segments of their IT infrastructure — their desktop fleets.
There are a number of forces coming into play here.
Perhaps the simplest of these is the traditional operating system generational change that we’re all used to by now. With the release of Windows 7 in late 2009, Microsoft has finally delivered a worthy successor to its ageing but still dominant Windows XP platform; and organisations right around Australia are eyeing off this long-awaited upgrade with relish.
However, the case for simply upgrading a bunch of legacy PCs to new hardware running Windows 7 is not as easy a decision as one might have assumed even two years ago. Today the whole paradigm of the desktop PC itself is under siege from a number of angles.
As Australian organisations start to reach the 100 percent server virtualisation mark (and Australia is one of the keenest adopters of this technology in the world), IT workers are looking around for other parts of their IT infrastructure to virtualise, and desktops are a logical next step.
Only this week, NSW electricity utility TransGrid emarked on a 50 user desktop virtualisation project, and it’s not alone. The National Broadband Network Company, the Department of Defence, the Department of Health and Ageing, and the City of Cockburn in Western Australia are all going down this virtualised desktop path.
These organisations are not playing around with thin clients. They are deadly serious about implementing the technology to drive real outcomes in their business. In the age-old centralise/decentralise IT paradigm debate, centralisation is in vogue.
However, there’s also more to the story.
PCs are highly adaptive tools which can be used for many different purposes. But if you’re only after a set of discrete use cases — such as entering data into a CRM platform and checking your email — many organisations are realising that a more specialised unit like an iPad may in fact be a better option.
Australian organisations like Brickworks are already close to ditching the desktop PC or laptop altogether for many staff, with sales staff increasingly equipped with iPads instead. “You can fully expect that as laptops come up for renewal, [the thinking will be] why would you buy a laptop if you could manage it with an iPad?” Brickworks business support manager Baden Bradbury said last month.
But wait, there’s more.
Desktop virtualisation is all well and good, but it’s still a ‘fat’ solution to a problem which can increasingly be solved through a ‘thin’ mechanism instead. With companies like Microsoft, Google, Salesforce and even Oracle making strong plays to provide software as a service through a web browser instead of via a desktop deployment, there is often very little reason why many staff members can’t do most of their work in a web browser.
I don’t expect Google’s pure cloud laptops running Chrome OS to take off in Australia, but what I do expect to see is organisations increasingly providing most of their corporate applications to their staff through a web browser or a virtualised application environment from a vendor like Citrix, and the employees themselves increasingly accessing those applications through their own personal PC, rather than a company-provided one.
Indeed, we’re already seeing this trend take force at innovative, technology-friendly companies like Suncorp, which has actually decided not to refresh its corporate desktop fleet. “People should use the device they feel the most productive in,” Suncorp CIO Jeff Smith told the Australian in late March. And we couldn’t agree more.
So what does this all add up to?
For small Australian organisations, the choice is becoming increasingly clear: Outsource as much as is safe to the cloud, facilitate your employees using the equipment they like best, and you’ll cut capital costs out of the business, as well as making yourself popular with your staff.
For medium and large organisations, however, putting together an internal business case around the future of a company’s desktop fleet is going to be a really tough ask in the short to medium term. This is a technology area which is in a constant state of flux at the moment, and few will be brave enough to predict precisely what will happen next.
After all, let’s not forget that the iPad only launched in Australia just over 12 months ago. It is basically impossible to say right now what the next 12 months will bring.
But one thing is clear. The days where Australian organisations would simply shell out hundreds of thousands to millions of dollars for a new batch of desktop PCs every three to five years are now over. We are in a new and more complex world now. And that’s not a bad thing.
Image credit: Bruno Cordioli, Creative Commons
Who buys desktops using capital dollars anyway? Most places I know just lease them (don’t even own the asset).
BYOD seems a bit like a vendor scam to me if you look at two scenarios under an MS environment:
1. LYDAH (Leave Your Device At Home) your company will lease/own a cheap desktop per user, have a CAL so they can do email/sharepoint/fileshare etc.
2. BYOD – end user buys a computer and a Windows license at consumer rates plus a maintenance contract. Company has to buy another Windows license to run their virtualised desktop, plus the CAL for email/sharepoint/fileshare etc. Instead of the cheap desktop – they have to buy high density blades, massive amounts of enterprise scale storage, upgrade the network, have extra data centre space, high end maintenance etc. All so some guy can come in and run a Citrix app or VDI app on their Mac or HP Pavilion? That’s all Suncorp provides – a virtualised CORPORATE locked down Windows desktop (not your own environment at all).
The latter seems a lot more expensive just so you can type on the same device you use at home at work.
This only seems to work if you are web cloud based corp – ie no requirement for MS, Citrix, VMWare etc
Hmm I take your point — and it’s a very good one. However, I seriously doubt the vendors are counting the BYO trend as a win … sure, they may come out ahead overall, but most of their staff would be incredibly frustrated after CIOs start telling them there’s no chance of the sort of massive, multi-tens of thousand seat refresh contracts that they’re used to.
I also think perhaps you’re overestimating the degree to which providing the back-office infrastructure is a big deal for the companies concerned … I think for many of them, it’s a logical (and quite cheap) extension of what they’re already doing.
I think you’ll find it’s a pretty big deal. From what I see – you can run 8 to maybe 50 VDI’s per average blade in a blade server – 6-8-10 blades per 2 RU. You need to step up the performance of the storage for those VDI’s too – and it’s not the same storage as you uses for your file store. And it’s the stupid things that kill the credibility – eg “lets move all the admin staff to VDI first they’ll be easy” – 2 weeks later “Jan the EA to the CEO can’t use their USB label maker or write a DVD, and John the communications guy runs an all company webcast and which kills the VDI environment”.
You’d be far better off ditching the virtualised “fat desktop client” in going to web – if you can take the hit on functionality.
Using an iPad is laughable – it’s ridiculously inefficient compared to having an actual keyboard, and laptops are still less productive than full sized keyboards.
Also, can you imagine the support nightmare of letting each person bring their own device? You’d lose heaps of man-hours trying to get them to work.
It was seriously considered here and just have everyone log in via RSA/Citrix. So you could either use your own device or a vanilla windows machine and log into work.
It was scrapped/plan altered.
We still have the RSA/Citrix thing’o but not the vanilla builds
Yeah I think a lot of companies are looking at it atm.
“Using an iPad is laughable” <-- depends what you use it for. If you use it for updating a CRM system on sales prospects and you're a door to door or a real estate agent ... it can be quicker and faster than using a laptop. And as for the support nightmare -- it's not really that much of a problem. If the machine breaks, after all, it's not the company's responsibility, it's the individual's -- and the company can always provide a standard PC for them to work on while they get it fixed.
I think the question of “what defines a computer” is going to change a lot over the next few years. Interesting to watch.
Truly — especially as we gain the ability to dock our iPads into corporate systems and temporary download virtual machines into them for the day’s work and that sort of stuff — that’s the kind of cool thing I am looking forward to ;)
Not really, the definition of a computer is pretty broad already and incorporates everything in use today. (http://en.wikipedia.org/wiki/Computer#Hardware)
the cheapest computer you can get is a $10 plastic calculator.
Ah yes, ToshP300 ;-)
You’re getting jipped if you’re paying $10 for a cheap calculator.
Some people may be able to replace their desktops with iPads and the like, but there is absolutely no chance my job could be done on an iPad.
heh or mine, my friend. Or mine.
Yeah but that’s only because Starcraft 2 doesn’t run on it.
:)
The only thing preventing me from using a vanilla iPad as my daily work device is lack of access to a serial port and multiple monitors. Jailbreaking fixes the first issue, perhaps they will do an IPad 3 with a thunderbolt port and fix the second.
Almost everything else I need is behind a web or ssh interface.
Google apps, dropbox and Xero FTW!
Desktop virtualisation becomes uneconomic for an enterprise with a Disaster Recovery site.
They will go from one desktop per person to multiple desktop infrastructure environments – one in the production data centre and one in the DR.
But you can run a VDI environment in active/active, and just have slower response / can lower priority (or lower SLA) machines in a disaster scenario… This is one of the ADVANTAGES of VDI…
An active/active configuration is not DR, it’s a distributed system. It reduces the need for DR but it is not DR.
DR is a cold, or at best warm, site that needs its own desktop virtualisation environment which is unused until a disaster occurs.
That’s the old paradigm – why have a whole datacentre sitting there doing nothing? With a properly virtualised environment everything is distributed and active, and the failure of one site just means that the less important workload is stopped until restoration.
Are you suggesting that a large enterprise that has a DR site (because it was the best paradigm at the time) should change over to an active/active configuation and also remediate all its old applications that can’t work in a highly availabile and virtualised just so it can adopt a virtualised desktop?
Spend a hugh amount of money to save a little? Try putting that in a business case and see what reaction you get.
New paradigms have to co-exist with old paradigms. If an enterprise has a DR site (the old paradigm) then the new paradigm of virtualised desktops is uneconomic.
Nope, I’m suggesting that an organisation moving to VDI, which has typically virtualised most of their server fleet, and is typically using a product like SRM to manage “DR” scenarios, could make use of the same technologies for their virtualised desktop fleet.
If the large enterprise already has a DR site for all of its legacy equipment it should be prepared to spend the money required to enable its key users to get back to work as soon as possible. There are many scenarios for DR that require a whole fleet of “backup” desktops – these also have quite some cost attached, after all.
Large enterprises are more likely to have more than one hub site anyway, so active/active still works.
Comments are closed.