- 11 September, 2008 13:39
The death of PCs?The death of PCs?
Virtual desktops are being touted as a way of delivering applications to any user via any device. So does this brave new mobile world mean curtains for the humble PC? ARN recently brought together a panel of industry professionals to hear their views.
Brian Corrigan, ARN (BC): What are the key benefits of virtual desktop technologies?
Rob Willis, Citrix (RW): When we talk to customers there are different drivers but the one where we are seeing the most rapid take-up is where people are doing work remotely or offshore and companies want to deliver the desktop securely. The other one we’re seeing is companies with a lot of desktops that are driven by cost, green IT or security concerns to do something differently.
Mark Mackaway, Datacom (MM): Everybody is testing the market at the moment so we’re seeing small sales within large companies that are trying to work out where it fits and assessing the hype versus reality.
Frank Mulcahy, Sonnet (FM): One of the big opportunities for virtual desktop is to smash the standard operating environment [SOE] management piece once and for all. Companies that have seen their fleet grow over the years want to simply manage a small number of SOEs rather than the many they’ve got today. Virtualisation allows you to split into layers so you have the base operating system with a separately managed suite of apps and local settings; to be able to do that and deploy it to any desktop simply is a big win for a lot of our clients. At the moment they’re just dipping their toes in the water because it’s a paradigm shift that requires a significant amount of capital investment.
Craig Taylor, Lucida (CT): We’ve got quite a lot of traction from customers that have gone down the thin client model with published desktop but found it isn’t suitable for many of their users for all the reasons why the thin client model wasn’t working. They’ve grabbed hold of this idea as a missing link from thin client through to their users. Having said that, we spend a lot of time trying to explain to customers what virtual desktop means because there’s still a lot of confusion.
Mark Little, Gen-i (ML): I think that’s our goal as independent integrators because we can go in and explain in real terms how it applies to particular clients, and how it fits into their BCP [business continuity planning] and DR [disaster recovery] strategies, as opposed to just focusing on the technology and its deployment options.
CT: A lot of them are also seeing it as a natural extension because they’ve been through all this server virtualisation and are comfortable with that concept so they’re ready for the next step. If you start conversations that way, you quickly find there’s some interest.
BC: Is Craig [Taylor] right to say thin client hasn’t worked and, if so, does desktop virtualisation offer companies like Wyse another crack at it?
Ward Nash, Wyse (WN): Thin clients account for 6 per cent of global desktops so you could ask why they’re such a small part of the market if they’re so great. It’s a question that has been asked for years but the difference was always that you couldn’t do on a thin client what you can do on a PC. If users couldn’t do what they wanted, then they didn’t want a thin client. What’s happened in the last nine months is that USB peripherals have allowed you to plug your Blackberry or iPod into a thin client. As we get closer to what a PC can do, we think that 6 per cent will go up. Will it ever be 100 per cent? Probably not but if it were 20 per cent I’d be super happy and driving a Ferrari. The difference as far as pilots go is that to get a whole bunch of applications to work in a shared services model was very difficult. If you wanted to get a pilot up and running, sometimes you had to port applications and get them to work together. With virtual desktop you can take a bunch of applications and stick them in a virtual machine – it might not do everything you want but you can certainly get something up and running very quickly. We’ve never been able to do that before. That means we’re able to go back into organisations and have another shot.
RW: It would be wrong to tie virtual desktops and thin clients too tightly together. While a lot of the work going on is around thin clients, the proliferation of different types of devices is another driver for virtual desktops because you can get it on an iPhone or put a Vista desktop on a Mac. Those are the things that are driving adoption.
Kevin McIsaac, IBRS, (KM): For clients I’ve spoken to, it’s not about moving to thin clients or even virtual desktops; it’s the idea of taking today’s image, which is an SOE, and getting away from that entirely to deploy operating systems separately from user data, user preferences and applications. If you move to that model, what lands on your desktop is essentially stateless. What matters then is how you tear the image apart so you can just run this stuff without having to build SOEs anymore. That means anybody can walk up to any PC in the organisation, login, authenticate and all those things come together on that device. That really is the key going forward and then whether you deliver that through a full desktop, an iPhone or a thin client. I think virtual desktop is a distraction and a step in the wrong direction.
FM: Another fundamental problem that hit thin clients early in the piece, and will also be a problem for virtual desktops, is the whole online situation. It suits some organisations but there are so many reasons why an online model simply doesn’t work. When you look at the capital investment needed to put the infrastructure in to allow it, and the real killer perception that you’re getting second best on a thin client where users are asked to take a hit for the team, it makes it a big step. Those are only perceptions, they’re not correct, but they exist in the industry. My personal view is that virtual desktop will suffer from that same perception as long as it remains an online model. If you can break the mould and take it offline then you’re really rocking and rolling.
Mark Gluckman, Regal IT (MG): The other thing you have to consider is that laptops are now $700 and getting cheaper all the time. Your biggest competitor is that people know they can get a new laptop; that’s your grey market.
KM: Laptop shipments exceeded desktops shipments globally for the first time this year and, by the way, smartphone shipments are about the same as laptop shipments so this whole idea that we’re going to move towards thin clients in a homogenous Microsoft type of world goes out of the window in the next 10 years. Instead, people are going to bring whatever device they want for a variety of reasons – change of generation being one of the big ones. Gen Ys have a preferred device and expect to be able to work on it; we are also using a lot more third parties in external contracts. It’s an Internet world that’s about standards.
MG: Operating systems are what we’re going to move away from but we’re still going to need to deal with disconnected users. People will want to work away from home without being connected to the Internet.
RW: I agree with what you’re saying, and laptops are really important, but I think there’s another point of view on that. The world is becoming more connected, not less connected, and everybody is running a 3G device of some sort. 3G is more than enough for a virtual desktop. Yes it’s true that people are becoming more mobile but they’re also becoming more connected.
MM: Children know that if they want an application they can go to the Internet rather than looking to see what’s installed on the PC.
MG: So why go to virtual desktops? Why not just run a browser and present it as an app?
MM: Because they’re home users rather than corporate users. We haven’t had any requests from a corporate entity to go to a Google model because there’s no personalisation.
KM: The [NSW] Department of Education has 1.2 million Gmail users. It’s my primary email client now. I was talking to one of the banks, which was looking to move part of its staff over to Gmail. I was horrified but you can’t say there’s no corporate interest.
RW: People are paranoid about security these days in corporate environments because they’ve seen so many people get burned. If the wrong email ends up in the wrong hands at the wrong point of time you’ve got another Enron. You can’t compare the security concerns of schoolchildren and the corporate world. If you go to a bank it has 1500 applications and the majority of those are homegrown. You can’t shift that to Google.
KM: The problem with virtual desktop is that it’s rooted in the old world of the desktop. In the next 10 years that will change dramatically – Google is one direction, I’m not saying it’s going to happen immediately but we are starting to see some movement. When you take the traditional image, break it up into its four layers and ask yourself how to do that, it has nothing to do with thin clients or virtual desktop. The question becomes what technologies I use to roll that out. When you take that strategic view of it, you have a really good place to start. If you talk to organisations with large, well managed full desktop, they’ve already got to the point where user data is separate and they’re starting to use application virtualisation to pull that out of the OS. The last puzzle in a full desktop is how quickly you can provision a basic OS onto a machine. About a third of my clients are halfway down that track already without going to the dramatic shift of a thin client or virtual desktop approach.
FM: [NSW] DET [Department of Education and Training] certainly changed my mind about how things get done. We were having a discussion about technology at home versus technology in the classroom and I couldn’t help but think ‘to Hell with what they’re doing in the classroom, look at what they’re doing in the schoolyard’. How are we going to be able to guarantee delivery of everything on demand to every device? There’s no one technology that can do that so you have to split it into individual components and use your best judgement.
KM: Device diversity is going to be a massive problem. Today the only solution is to go to an SOE model.
RW: I disagree because we [Citrix] don’t want to be specific to anything other than the ability to deliver something on a screen. You can go to YouTube and we’ve demonstrated virtual desktop on an iPhone. We’ll run it on anything – [Blackberry] Bolds, Linux, Macs – to provide an XP or Vista desktop anywhere on anything.
KM: The problem is that applications are tailored to a large screen, a mouse and a keyboard. As soon as you put that on a Blackberry Bold you are scrolling around to get things into the screen.
MM: It’s going to be a long time before one of these [smartphones] is a corporate device that a company will be happy for their staff to sit down and work on because they’re just not user-friendly and you can’t be productive enough.
RW: There’s a time and a place, such as sitting in an airport lounge, where a smartphone screen is sufficient. It isn’t what you would use eight hours a day but if you are mobile and need to get something done it works fine. It’s that ability to go from one to the other that people want.
FM: I agree with one addition. Every user has a work profile and that will dictate what’s best for them at any particular time. The one thing I would add though is that the user was asked to compromise with thin clients and they’re no longer prepared to do that. If they’ve chosen a device, they want you to deliver to it. It’s on demand and if I choose to use it on a mobile device then it’s my choice but I won’t accept anything less than full functionality on my desktop.
Dipping a toe in the waterDipping a toe in the water
While attendees were keeping their cards close to the chest, several integrators at ARN’s recent virtual desktop round table said their clients were piloting virtual desktop environments.
Datacom’s Mark Mackaway pointed to the benefits of centralisation as a key driver.
“Grunty applications tend to have large back ends that require support and it’s usually a lot quicker to get data from the back end to the virtual desktop,” he said. “We have customers looking at blade workstations for that extra level of power and control but it’s still managed under the virtual desktop broker and distributed under the same mechanism.”
Extending that point, Sonnet’s Frank Mulcahy said the virtual desktop model allowed greater flexibility because a user that needed more power could have it supplied on demand from the datacentre.
Wyse’s Ward Nash pointed to Flight Centre, which has 27,000 thin clients connected to Citrix, as a leading example of how market perception was changing but said banks and government departments were still reluctant to say what they’re doing. Sonnet’s Mulcahy suggested banks’ reluctance to talk about virtual desktop deployments was largely because this would force them to admit they were using it to go offshore but IBRS analyst, Kevin McIsaac, said they were still using it exclusively in very small pockets.
“It’s a niche application for security, temporary desktops and occasionally six-star buildings transferring heat into the datacentre but it does not reduce overall heat footprint,” he said. “The whole thing about reduced cost is a furfy because costs are roughly the same as a well-managed, full desktop environment. There’s a massive hit upfront and you may or may not get savings 3-4 years down the track – it’s a huge risk.”
Citrix’s Toby Knight said one of the biggest challenges large customers were trying to solve around their desktop environments was around moving, adding and changing desktops. He cited one company that had 8,000 users and made the same number of move, add changes in a year.
“They’ve got a well-managed environment with the best system management tools in place and it still takes them 4-6 hours to do a move, add, change today. It doesn’t happen instantaneously,” he said. “The second biggest thing we’re looking at – and I’m talking to two or three leading organisations in Australia – about moving 30 per cent of their workforce outside of the four walls and tapping into the skills of user groups like stay-at-home mums. Using a traditional delivery mechanism around a distributed PC, they just can’t get access to those people. It also reduces the amount of property and other overheads.
“We are working with these guys on niche deployments today to prove the technology but the conversations we’re having are about broad adoption. The technology still needs to mature before it goes mainstream but it’s not a six-month conversation; it’s a 2-3 year conversation.”
The road to mainstreamThe road to mainstream
So what needs to happen before virtual desktops become a mainstream technology? Citrix’s Rob Willis argued it was simply a matter of time.
“All the drivers are there, and the underlying technology is there, but you don’t go mainstream overnight – especially with desktops because there are so many moving parts,” he said. “Server virtualisation went more quickly from something that was interesting to mainstream adoption but there aren’t so many moving parts. In the desktop there are lots of things you have to put together, and user experience will play a large role in speed of adoption, but it will happen because there are too many drivers.”
So where is it happening? Willis said the US, Europe, Australia and Japan were the key geographies so far. While US banks might be expected to take a lead, he said that hadn’t been the case so far because the credit crunch had negatively impacted deployments.
“These are complicated issues that touch guys in lots of different parts of the business so it takes time,” he said.
Datacom’s Mark Mackaway pointed out that he had been virtualising desktops for customers since 2002-03, though admittedly in small projects, so the technology was not new. The difference now, he claimed, was that we are starting to get the technology to meet client expectations. While there was a common perception that server virtualisation had caught fire very quickly, IBRS analyst Kevin McIsaac said it had taken longer than people think because the products were ready to go in 2002.
“It wasn’t really until late 2006, early 2007 that is started to take off and even now only 30 per cent of servers that are shipped are virtualised,” he said. “That was a 5-6 year period and if this [desktop virtualisation] is going to take even longer, we could be looking at 10-12 years. “I’m not saying virtual desktops don’t have a place but they’ve been massively overhyped in the past 18 months and it will be very, very slow.”
Regal IT’s Mark Gluckman said some smaller virtual desktop environments would grow organically but that the market desperately needed a major organisation to take the leap and make a large, upfront capital investment. As yet, he was still to see one.
Sonnet’s Frank Mulcahy said there was always a cost discussion with senior business managers but that it was generally preceded by a risk discussion.
“They’ll say ‘everything is working, it’s not great but it is working. Now you’re asking me to take a punt on something’. VMware was banished to test and development for 10 years then all of a sudden it gained mainstream support and away it went,” he said. “With desktop you’re not putting all your eggs in one basket but if you take a hit you’re down 20-30 seats in the dark and might as well have put adequate failover in place. It’s a risk and at the moment and users don’t have that risk. They want to hear a very compelling financial story and there generally isn’t one. “We’ve all been in the industry for a long time and had these discussions when thin clients first came out but we had wins then because we found clients that didn’t perceive it as a huge risk, were very compelled by the financial story and really wanted those technology gains. “That’s where we are today [with desktop virtualisation]. It’s not mainstream and there’s an element of risk, but give it some time and people will accept it.”
Creating a marketCreating a market
ICE Systems’ Gary Spooner asked whether there was anything else the vendors should be doing to help their resellers get opportunities on the table.
“We need to be able to present to our customers but their users seem to be the other firewall so we have to help customers put propositions to their users,” he said. “Customers want to stay in business and make more money so they need to drive economies of scale back through the organisation and users need to find some way to accept it.”
Citrix’s Rob Willis has done more than 120 demonstrations of virtual desktop technologies ranging from 5-6 people to more than 100 at a time and suggested the market was over the hump.
“We have a lot of activity within large customers looking at it for different reasons. I’m not surprise they’re not making decisions around 20,000 desktops in one go. To me they’re following a completely normal process,” he said. “I would agree that vendors are leading the communication around what it is and how it is but it isn’t simple so there’s a level of commitment needed.”
He claimed the response to recent partner training sessions had been incredible and that adoption was where it should be at this stage.
Anything being labelled as the ‘second wave of virtualisation’ is sure to create interest as the industry looks for its next big opportunity. itX’s Greg Newham said partners were keen to come onboard because there was so much hype around virtual desktop but that it had yet to translate into significant business.
“Certainly the information sessions we’ve run have always been well attended but the rubber still isn’t hitting the road in terms of good case studies or compelling reasons why users are going to fall over themselves,” he said. “We’re obviously keen to see this translate into real business but at the moment there’s a lot of noise without significant dollars being generated.
“You can’t have this much energy going into a market without producing anything but we’re still priming the pump at this stage.”
So where will the industry land large deployments that can be turned into case studies and drive broader adoption? Perhaps surprisingly, IBRS analyst Kevin McIsaac suggested banks could be the best bet if they can get it to work adequately.
“If you look at who has been talking about it in the media during the past 12-18 months, it has usually been the financial institutions,” he said. “It’s usually one individual that makes a big bet. The banks haven’t actually found a use for it beyond very specific examples and I think the strategy, if you want to be successful in this market, is to give up one the idea that this is a general technology to replace the desktop and chase those niches first.”
Datacom’s Mark Mackaway said he knew of at least three major organisations that were looking at ‘boots and all’ deployments but predicted they would not allow anything that does happen to be used as case study material.
ICE Systems’ Spooner asked why the major vendors didn’t sweeten the financial story on a couple of potential major contracts to give the market a kick start.
Citrix’s Willis said there were some in the works, including a major international insurance company, and one of the big four banks that was gradually growing its footprint.
“You start off having a conversation about virtual desktops and end up talking about eradicating standard SOEs. That’s really where you start getting transaction,” he said. “If you can’t have that conversation, you’ll die on the vine but those that can make the transition will be on a winner.
“In our large customers we’ve been able to make that transition away from what virtual desktop is today to what it will be in the future.”
Is there a need for standards?Is there a need for standards?
To illustrate one of the biggest problems with virtual desktop today, Datacom’s Mark Mackaway drew a comparison to the battle between Blu-ray and HD-DVD.
“Some people are saying it’s one or the other [Citrix or VMware] when we really need to get the complexity across. There are so many products that you can pick or choose,” he said. “Interoperability creates issues in some instances but that’s where we as integrators can talk to clients about what they are and tailor products so we have a suite that works in any given clients environment.
“That generic deployment experience is one of the key things lacking from any of the integrators. That’s not to say the skills aren’t there – we just need to avoid the polarisation argument because we are at the steep end of the learning curve.”
Regal IT’s Mark Gluckman was encouraged that key vendors were starting to talk the same language and said that would eventually get through to the user community.
Citrix’s Rob Willis said there was already a great deal of standardisation but that it was “under the covers a lot of the time”. However, he said there would always be differentiation between vendors and that was good for the market because it provided choice for users and partners.
Sonnet’s Frank Mulcahy said one of the advantages of being in a two-horse race was that you could call your technologies standard because nobody could really argue with you.
“Citrix as a company, if it wants to help the community, needs to define its standards very clearly around its own product set,” he said. “We need to simplify the message because it’s the people who aren’t involved that we really need to get to.”