Death of the business Desktop
This topic was created by Chris Mellor 1 .
Death of the business Desktop
The business PC desktop is facing death by a thousand VDI cuts augmented by a BYOD bashing.</p>
<p>Business desktop death pointers:</p>
- EMC’s XtremIO all-flash array, twinned with VMware View or Citrix Xen Desktop and backend VNX arrays, can support 7,000 or more virtual desktops
- All-flash SolidFire arrays can support large-scale, 1,000+ VDI roll-outs
- Atlantis and all-flash Violin arrays can support 1000s of virtual desktops
- Hybrid flash/disk array startup Tintri can support 1,000s of virtual desktops
- Startup Pivot3 supports VDI
- There is a VDI-focussed Vblock from VCE
I know this is repetitious; that’s the point. These aren’t just a few pointers; this is a flood, a veritable tidal wave of systems all focussed on removing pricy and complex-to-manage business desktops with centralised virtual desktop systems.
Some other suppliers with VDI capabilities; Pure Storage, Tegile, Nimble Storage, and Fusion-io. A combination of flash storage and deduplication is making it possible for cost-efficient and storage capacity-efficient VDI set-ups with the responsiveness of actual desktops, or better.
Set this VDI blitzkrieg to one side and consider BYOD - Bring Your Own Device, in which users bring their own notebook computers to the office. This is the guerilla war assaulting the business PC with VDI being a full-on, frontal assault.
The net result could be a multi-year reduction in business PC use with, in some businesses, desktop PCS literally disappearing.
We haven’t any numbers, beyond the general PC annual shipment numbers decline. Our sense of it is that the business desktop is an endangered IT species facing year-on-year shipment declines, wilting under the impact of artillery barrages from the massed ranks of VDI howitzers and BYOD sharpshooters.
In that case there will be a consequent decline in business PC component shipments; hard disk drives, power supplies, DRAM, motherboards and CPUs.
If 200,000 business desktops go away each year for five years that’s a million fewer hard drives shipped. And it could be worse; 500,000 fewer desktops each year means 2.5 million fewer drives over five years.
I think we’re at a VDI/BYOD tipping point and a storm surge of virtual desktop instances is going to wash increasingly unwanted and unloved business desktops right out of the offices they’re anchored to, never to return.
Is this true? Will it happen? Am I smoking pot?
I think not Sherlock.
Re: Death of the business Desktop
So let me get this right. The death of the PC is because XtremeIO can host 7000 virtual dekstops and others can host thousands?
Mr Potts has it pegged.
Re: Death of the business Desktop
Article an obvious troll for comments. Okay I'll bite. Yes, virtualization brings benefits but it also increases complexity. All of the old abstractions still exist - server, disk, NIC, OS, memory, swap, switch, vlan, etc., etc., just now they are wrapped inside another set of abstractions. Staff need (much) more training, not less.
Meanwhile vendors throw a few more skins on the onion and say they are saving you money. The hapless Solaris admin was using "format" to partition disks in 1995. Now he is using "format" to partition disks in 2013. Only now you have to send him on a series of expensive VMware courses so he can provision the virtual disks before using "format" to partition...
Re: Death of the business Desktop
No VDI isn't the death of the business desktop, it might be the death of the fat-client desktop model we've been accustomed to for the last 20+ years as thin-client finally makes the grade. But thin-client is still a business desktop...
Original text the below quote is here.
If System Center licensing required a spreadsheet and a cheat-sheet, to truly understanding VDI you’re going to need a young priest, an old priest and possibly a bottle or six of scotch. To start with, VDI has a few different flavours.
The first is Remote Desktop Services (RDS) - otherwise known as Terminal Server - in which multiple users log into a single Windows Server instance. With the new Fair Share feature and a lot of the enhancements to RemoteFX in RDS, this is a method worth considering if only for the simplicity of it all. You licence the appropriate number of Windows Server virtual instances to host the number of users you require, you buy a Windows Server CAL and an RDS CAL for each user and you’re done.
The extension to this is that you can give each user their own virtual machine by simply buying a Windows Server Datacenter license and spinning up one Server VM per user. This method was popularly considered a neat way to sidestep the byzantine complexity that Microsoft’s Vista/Windows-7-era VDI licensing was, as WS2012 allows two administrators to connect to the operating system without needing RDS installed.
Microsoft has explicitly forbidden this approach; if you are using instances of WS2012 as individual VMs for your users, you must get an RDS CAL for each user. Unlike the traditional RDS approach, however, your users can have their own VMs.
Microsoft’s official stance is that VDI is best done by using a Windows Client operating system. This is where things get dark. The intuitive licensing approach - buying a copy of Windows XP, 7, or 8, installing it in a VM and letting your users have at it - is explicitly outside of compliance. As soon as you put Windows Client inside a virtual machine - or remotely access a physical machine through RDP - you must licence the client device as well.
There are some basics to VDI licensing for Windows Client operating systems to know. Microsoft offers a licence called Virtual Device Access (VDA) which costs $100 per year. It is available only through volume licensing; signing a volume licensing agreement means agreeing to have Microsoft audit you each year, so make sure you have your ducks in a row. VDA licenses apply to client devices. Windows endpoints covered under Software Assurance (SA) have VDA rights and so do not need to pay this fee.
A single Windows VDA license allows that endpoint to be RDPed into a maximum of four Windows Client instances simultaneously. You must hold as many VDA licenses as you have thin clients and non-SA-covered endpoints accessing the VDI environment.
Extended Roaming Rights are another critical concept. Any device covered under VDA or SA allows the user to access a Windows Client from up to four x86 devices, so long as those devices are outside the corporate firewall. This is designed to allow users to access their work PC from home, or a personal laptop.
If that personal device - the canonical example is the personal Macbook - is brought into the office, then it is considered part of the work environment and must be covered by VDA or SA.
To summarise: if you have a PC at work covered by SA or VDA then you can access up to 4 VDI instances of a Windows Client operating system. You can use up to four computers not owned by the company outside of the corporate premises to access those same VDI instances. If you stop using your Macbook at the Starbucks across from the office, walk into the office with it and use it to RDP into a work VM, then that Macbook must be covered by VDA or SA.
ARM devices are licensed differently. If your company covers its PCs with Software Assurance and buys you a Windows RT tablet, then you can access a Windows Client instance from that Windows RT device.
Corporate-owned iOS, Android, Blackberry or other ARM devices must purchase a Windows Companion Subscription Licence. CSLs are per-user and cover up to four companion devices. Personally-owned ARM devices - including Windows RT devices - must be licensed under a CSL to access a Windows Client operating system.
None of the above brings Intune or Office 365 into the discussion, considers diskless PCs, or attempts to explain licensing Office in a virtual environment. It is not intended to offer exacting guidance: VDI licensing can change at any moment, and making sure you get the best possible set of licenses for your deployment can involve factors beyond those discussed above. It is always best to discuss your VDI plans with your VAR or Microsoft rep.
If possible, try to get a stamp of approval from MS on your deployment before you deploy. Remember that the burden of auditing is on the business - not Microsoft - and you may be asked to verify your compliance every year. It is strongly recommended that anyone considering VDI invest heavily in automated compliance tools. You’ll need them.
The long and short for me: I clocked my access in 2012 to my home XP VM at just over 300 devices. At $100/device/month for "remote usage rights" Microsoft would have me pay $30K/year to access my home VM. Of course, I can't access my home VM because I can't get remote usage rights without SA. I can't have SA without being a company. So I the initial cost for just that one licence is up a few thousand as well.
If I want to stand up the ability to allow my staff to access Windows 7 instances from any device anywhere in the world it is going to cost, and cost and cost. Every new endpoint that connects is $100/device/year, and Office is yet more.
Microsoft says "just use Server + TS and use shims to fool apps." Sadly, I have a number of apps that won't work on Server and for which shims have not been - and likely never will be - written. Microsoft's response is "tough titties." (That's not even getting into a Server licence + CALs, etc is a hell of an entry-level buy in for someone to access a home VM, or an SMB worker to use a centrally provided desktop on the road.)
For me, I am dumping every last spare dollar into both supporting Weyland/Weston with the new FreeRDP compositor and porting my apps to Linux (for graphics intensive stuff) and standards/HTML delivery (for everything else.) Until Weyland/Weston/FreeRDP are ready to rock on RHEL/CentOS (a few years yet) then VDI just isn't something we can afford.
Your mileage will vary.
We are the Beast of Redmond.
Lower your expectations and surrender your productivity.
You will use both biological and technological means to accomplish what should doable with technology alone (we won't licence the technology to the likes of you.)
Your business will adapt to service us.
Resistance is futile.
So it's all messed up due to licensing?
I am sure they can change that.
Outstanding analysis, thanks.
Beautiful summary of the VDI licence nightmare. I've been trying to work out the best scheme to provide VDI for students. There are two ideas, one is to provide Students Access to a "virtual" version of our Lab PCs to access from Halls and Home during holidays. The other is provide remote students a desktop to run some specific applications.
Microsoft licences just make this impossible to manage. A typical student has at least a laptop and a phone, and likely a (non-Redmond) tablet, and probably their parents desktop? - so VDA needed (and compliance managed!) for each device they want to access. Oh, and they'll bring the laptop to the Campus and try and do the same.
I have a feeling the only way to make this work is to loan each student a Surface RT and restrict access to those!!
Can the Reg get hold of whoever came up with this stupid scheme and make them explain themselves (or resign from Microsoft). We just want to buy an OS licence for a VM and share it like a normal student PC!!! ARRRRRRGGGGGGGHHHHHH!
If I could get my hands on the poxy whoresons responsible for this spectacular clusterfuck "interviewing" these wretched accretions of mewling fundamental evil would be the singularly last thing that came to mind.
This is a nonsense article. The cost of thin client devices is still pretty high (especially if you want pass through graphics, any sort of caching or 3d acceleration etc etc..) .. combined with the quite ludicrously expensive and complicated licensing required for VDI implementations (including datacentre server licenses and all sorts) means that it's cheaper by far (over a 5 year desktop lifecycle) to buy a desktop.. and that won't change I think any time soon until MS themselves offer custom VDI instances to corporations in the cloud at some standard rate or other.. then it makes a ton of sense for the majority of office based "vanilla" users.
Although there are all sorts of just cool as hell hardware technologies.. the management of the clients, the running of all software successfully in the virtual world and especially the licensing on top.. means it’s pretty evil...
Unless you are a Walmart type business model or a university and so the vast majority of your people can happily live with 3 functions provided via a cheap as chips terminal device.
But Trevor, tell us how you *really* feel. :)
(I expect that capturing the carnage on video for that 'interview' would be something that you could sell for a nice amount, though. )
Unless you are a Walmart type business model or a university and so the vast majority of your people can happily live with 3 functions provided via a cheap as chips terminal device.
There are always tiers of users within organizations, with differeing requirements. Where VDI breaks down is with mobile computing. For large businesses, it tends to be more senior execs and their staffs who make use of these toys. The rest of the users tend to have no say in what they use and will thus end up with the cheapest kit around. Thus: tablets and laptops for execs and VDI for the masses. I predict this will hold especially true for government agencies.
Licensing schemes will change to fit the VDI model, or the people selling the licensed products will lose business. This article is about where things are headed bases on the present as a starting point, not that we are there now. For large corporations, costs of management will be brought down significantly as there will be much less need for touch labor. Of course, VDI implementations will reveal new problems and reinvent old (i.e. poorly managed permission sets), but that does not mean it is unlikely to catch on. It already is (and I am having to deal with it).
I can't tell you how I really feel. I still occasionally have to cross the border into the US and the NSA monitors everything. Free speech is dead. Fear big brother, because I sure as fuck do.
Larry Ellison must be turning in his grave......
Re: Good God
What? Is he a vampire?
Re: Good God
"What? Is he a vampire?"
Well, he represents one. Perhaps he's a class of familiar.
Not on this desktop
When it came time to replace my 'trusty' (hangy) work laptop I opted for the Xeon workstation instead. My excuse was that it was cheaper.
That said, I think in many shops there is a glimmer of realization, as they get more comfortable with managing a lot of VMs, or Citrix boxes, or whatever, that the biggest (though not only) obstacle to VDI adoption may be Microsoft Corporation.
The unasked question is whether we can also replace or commoditize laptops. The answer is not as obvious as for desktops. Sales people will fall into 3G black holes. Execs will want local USB for their valuable presentations. Execs also want their desktop to be at least on the same continent. Fat client. (Don't say this out loud.)
For anyone chained to a desk or who has no rights, VDI looks great, if you can make Microsoft go away.
The RDP experience, when it's well done, is excellent. You do not need fat networks to the client. However all video acceleration is done locally. This has interesting implications for the economics of Intel v AMD v ARM at the client.
Re: I would imagine...
Whaaaaattt ?? £600 a year per user ?? That's 2 desktops a year.
Make it £5/mth, or maybe £10, and I might bite.
Licensing is the killer
As the chap above said - even if you want to roll out a thin/zero client VDI infrastructure with a 1:1 mapping between thin clients and VMs, you have to pay $100 per endpoint, per year - you can't get SA on a thin client that isn't running Windows.
So a half-decent thin client costs $300, then you have to pay $100 a year - so over 5 years, that's $800. And you could buy an equivalent fully licensed desktop for $600. So to the beancounters you're having to justify the additional expense of $200 per endpoint, on top of the storage, servers, and hypervisor licensing to run the back-end, which is, realistically, another couple of hunderd dollars per VM.
On top of that, someone who does standard desktop support and deployment probably doesn't have the first clue about managing a virtual infrastructure, so you either have to spend on consultants, send your desktop team on training courses, or hire someone with the right skills, which actually increases your spend on that aspect as well.
Where this works well is getting the economy of scale on the back-end and have someone else manage all that tin for you. Unfortunately, Microsoft also expressly forbids using the same servers OR SAN for different customers in a VDI deployment (presumably to stop people chucking their desktops into EC2).
Funny how things go
Seems pretty clear that the next step is binning those thousands of virtualised DesktopOS instances and replacing them with a server/cloud running a multi user OS that can stream apps and data tweaked for whatever device we happen to have to hand, which sounds a bit like what Google's aiming at with ChromeOS and their NaCl runtimes.
As noted above, licensing is the real issue for VDI. So much so that VMware are now supporting 2008 as a desktop OS just so service providers can get around MS's absolutely archaic licensing. The improvement of storage is definitely one of the reasons why VDI is taking off more (don't forget stuff like Atlantis, Pernix, CBRC being in View and even the ability to use local SSDs for non-persistent desktops in both View and XenDesktop now). The other main reason why VDI will be considered more over the next 18 months is that we FINALLY have support for decent virtualised graphics. vGPU is freaking awesome from what I have played with so far.
At the end of the day VDI is a stop gap anyway, the holy grail is to have all applications virtualised, we are just a long way from that. In any case, MS will continue to try to cripple VDI sales with their licensing until RDS can compete with XenDesktop and View.
Nice trend, crappy licensing
I've trying to make VDI projects happen to many customers, but all but a few have come close to fruition. ROI and TCO projections are always the problem, even if you take into consideration operating and administration costs for 5 years.
Infrastructure costs, especially storage, are really high and that's because sizing guides provided by VMware, Citrix and co always underestimate the IOPS/Throughput and read/write ratio for VDI data.
New flashy arrays such as ExtremIO and RamSan/IBM FlashSystem will help bring the storage cost down. I've tested the new IBM Flashsystem 810 and it's absolutely awesome, but this takes care of only half the problem.
Software solution like Atlantis iLIO are quite interesting but the bring fundamental changes to the usual VDI model and still have to find traction in the market. Mokafive seemed to be an interesting approach but any attempt to contact them has ended in radio silence on their behalf.
Big companies are tied to Microsoft and migrating to a VDI or DaaS environment implies enormous licensing costs as MS tries to ride the VDI wagon as profitably as possible.
Desktop performance and user expectation are big issues that need to be set in the early stages of a project.
A basic 500 EUR PC comes with Core i3, 4GB RAM, 500GB HDD, Win 7 Pro license and a decent 19-20" display and it can handle most workloads a usual company employee can throw at it. Trying to match the same level of performance in a VDI environment requires careful user profiling and monitoring and for relatively normal use cases you will find yourself proposing zero client solutions using the Teradici chipset and VMware Horizon backend, both of witch come with a higher cost per user when you factor in Windows licensing costs.
And then you have the peripheral devices problem. Specific devices such as USB scanners are a nightmare to integrate in a zero client VDI solution.
Last but not least if the end user has a distributed environment, with many small locations you will find yourself modifying the networking infrastructure as to ensure at least a basic level of functionality if the main connection to the central datacenter should fail.
Another thing to consider is user rejection of the platform, as it can be viewed as an advanced form of employer control.
All in all I think VDI, BYOD and DaaS are the next big thing in corporate environments but VMware, Citrix and Microsoft need to come to their senses and bring licensing costs down or at least provide an alternative licensing model. I guess a monthly per user fee would fare much better. The other alternative would be a completely open source solution based on KVM and maybe RedHat would be a great alternative if they could provide a stable, scalable and cheap solution.
Re: Nice trend, crappy licensing
"ROI and TCO projections are always the problem"
Err, so apart from the fact that the financials don't work, this is an excellent solution.
Remind me again why VDI is meant to be such a great idea ?? Surely it isn't just being punted at the hapless users by the vendors looking to make increased profits ?? Never !!!!!
is licensing a stalling tactic...
It's been said in the past, and it would seem to continue to be the case, that Microsoft are impeding desktop virtualisation until it suits them.
Somewhere else (I think on The Register) there was an article about technology companies heading down the "devices and services" route that MIcrosoft are most visibly doing; notably, Oracle, IBM and Google.
- Microsoft has Nokia
- a Nexus 5 phone class of device costing £300 has more computing power (at a hardware level, at least) than a desktop of just a few years ago
- RDP has and continues to be refined for more use cases
- WAN friendly
- clients for iOS, OS X and Android
- Windows Server 2012 R2 introduces
- production-grade virtualisation
- storage tiering
- "shared nothing" virtualisation
- Windows 8.1 has been shoehorned into an 8" tablet costing just £249 with 8 hour battery life
- Microsoft continue to shepherd users to the cloud, with improving (capability-wise, cost-effectiveness-wise, but perhaps not yet availability wise) cloud services
- Microsoft have announced plans to provide Enterprise Voice (ie a real phone number) in Lync Online from Summer 2014
It would seem Microsoft are marshalling a scenario where
- you buy/subscribe to a device, much like you buy/subscribe for a mobile
- renew every 18-36 months, to keep the revenues coming
- you connect to a [leased] VDI instance (if necessary) from your device
- to keep the revenues coming
- use Miracast (for wireless display) and Bluetooth (for wireless keyboard and mouse) when you need a "conventional" keyboard/mouse/monitor interface
- your device is also your unified communications device, including telephony
Are corporate IT being reduced to "plumbing" - wired and wireless networking...?
Re: is licensing a stalling tactic...
People better hope that Linux, probably RedHat, can develop this and businesses will need to be willing to break their Microsoft addiction cause once MS has all this down and your business on it. Well wow, they will own you.
Post PC era
I've been hearing about the so-called "Death of the PC" since 1996 when Scott McNealy confidently proclaimed that his Java Station thin client was suppose to replace PC's. Problem is that thin clients and flash arrays are dependent on company's spending more money on Servers and Server management than on desktop computers and server management combined. We're not just talking about flash memory arrays, we're talking storage, and since at this time, Flash memory is much more expensive than HDD disc storage, this "Post PC era" hasn't happened and isn't going to happen any time soon. Bean counters, are the reason why. There is no cost to performance benefit in eliminating desktop computers. Desktop computers are cheap and aren't complex to manage because they have their own memory, storage, and CPU and therefore, are not dependent on a server for memory, storage, and CPU. They are only dependent on networked servers for networked connections and the internet access, e-mail, and in any networked software or additional services. And if, as in my day job, you work in a call center doing tech support for Enterprise Mobile Device Management and all of the system that you use run inside of a browser, you sure as hell wouldn't want your computing entirely running from expensive servers that would have to be deployed in greater numbers than your desktop computers, just to allow you to do your job.
Perhaps when the cost of Flash Memory reaches the cost of HDD disc storage, we can begin to think about this Post PC era situation, but until thin, while small businesses with more money than sense, could benefit from a "network computer" thin client situation, larger enterprise businesses most definitely would be wasting money in eliminating desktop computers from the work place. There simply is no cost to performance benefit in doing so, at this time.
Re: Post PC era
Uh 1000 physical PCs for a call centre are easy to manage than 8 servers and running everything through something like XenDesktop and RES? You are kidding right? MS licensing might be a b*tch but if you are doing non-persistent VDI then you can definitely get to a similar cost as a physical PC deployment with a LOT of added benefits (easier image and application management, hot desking, remote connectivity). Large enterprises is exactly where VDI plays its biggest strengths.
It is not about cost to performance benefit, it's about cost to management benefit. You just need to make sure that the end user experience is the same level as on their PC, which is doable. VDI projects fail or are too expensive at the moment because of bad consultants/resellers. Not because of the cost of the tin or the tech itself. And flash storage does not work out more expensive in the VDI world than disks. Maybe if you are looking at having an all flash array or putting SSDs in something like a 3PAR but in that case you are doing it wrong. Use cheap local SSDs or something like Pure that has good dedup (seeing as most VDI data is identical). You don't need capacity, just enough performance.
the Linux world (for comparison)
In the world of Linux and Free software it's been a non-issue for a long while. You can run stuff locally. Or you can run stuff on a remote server with X display across the network. And there's actually not much difference between a thin client and a desktop PC ... they're both systems running Linux.
The key hardware issue is network bandwidth. In theory a Gbit LAN can transfer data at 120Mbytes/second, which is about the same as a hard drive (but considerably inferior to an SSD). In practice there are bottlenecks. As for network X-serving, the difference in bandwidth is far more noticeable if (and only if) your work necessarily involves interaction with moving imagery (especially interaction with 3D models). The CPU to GPU / Graphics card bandwidth is enormous compared even to a dedicated flat-out Gbit LAN.
But most office-pushers could be on thin clients were it not for the baroque tangle of Windows licensing. Go Linux + Libreoffice on thin clients and a fat server and LAN!
Corporate of course
I can see VDI removing the desktop for large and enterprise customers for sure. SMB might do the same but with BYOD programs. For me though a PC will always be a necessity just for PC gaming. Better looking than consoles, mouse/keyboard is far superior to gamepad as a controller, and being able to do so much more with a PC than with a console such as multi-tasking.
I've participated in roll-outs of VDI in concert with BYOD programs and when done right the cost savings and better UX are definitely there.
A funny thing, well not
Not so long ago 2010, I with the help of others, came up with a VDI solution for a client involving Sun, VMWare, Citrix (Legacy) and Microsoft, in some form or other. The main driver was security, no documents on desktop computers, no USB ports, and the ability for separate corporate networks to remain separate, but visible through a single terminal device, for users allowed to access more than one network. Only one solution supported that, Sun Ray, all was going fine until Oracle took over Sun and started insisting you used Oracle VMs. We actually wanted to mix the back ends, some Windows Server Based (External), some VMWare (internal). The biggest headache was external users, maybe 100 on a bad day at most, but out of a pool of maybe 10,000, on 100,000s of endpoints worldwide, basically the external bit could run on a single 2U 2 Socket server. VDI was the answer, the technology would work, but the licence costs were ridiculous for external users. Guess which vendor couldn't get their head round the issue. Outside of licensing the other problems we had were Server sizing, bandwidth and latency which were far more than any of the vendors predicted, and application compatibility. The basic office suite, and a few common well behaved apps worked fine, but once you try the specialist and older client/server applications, you have a problem.
So, why would you want to use VDI on a browser driven application, once you have everything browser driven, you don't need VDI. Heavy weight desktop apps maybe, but if it requires a big PC then a big server will be required to host it. Some terminal services client can operate within a browser very effectively.
Life would have been simpler if MS had let us use our own Office 365 implementation, and we'd have ended up with a server farm a fifth of the size of the VDI one required.
"I know this is repetitious – that’s partly the point. These aren’t just a few pointers, this is a flood, a veritable tidal wave of systems all focused on removing pricey and complex-to-manage business desktops with centralised virtual desktop systems."
What is it with humans? Everything goes round and round in circles, centralised, de-centralised, centralised, de-centralised.
MAKE YOUR FKING MIND UP and STOP CHANGING EVERYTHING EVERY FKING 5 MINS!
I can't even imagine the nightmare of replacing our desktops with a VDI solution. It might make sense for an organization with a few hundred desktops in one location, but for us, with 20 locations spread all over town sharing 2 server rooms connected by a fiber optic MAN and somewhere around 7000 desktops, it wouldn't. It also doesn't make sense for the small businesses that make up 80% of all businesses in the US and have, at most, a dozen or so computers.
So basically, no, the business desktop is not going away anytime soon. If anything were going to replace it laptops would have (and, to be fair, have for a lot of businesses, but not enough to kill the market).
With regards to "It also doesn't make sense for the small businesses that make up 80% of all businesses in the US and hve, at most, a dozen or so computers"; wouldn't the owner-managers (assumed) of these businesses be fed up of the headaches of IT management, and happy to rent it all from Microsoft - Office 365, Windows Phone, rent VDI instances powered by Windows 8.1, now memory-optimised for VDI, disk I/O optimised for VDI, RDP optimised for WAN, integrated A/V. After all, how many businesses *own* their vehicles, and how many are happy just to lease them? In fact, a sealed, limited but "maintenance free" appliance (a bit like a Windows RT device) would be a great rented VDI client...
Don't have any answers, just playing devil's advocate :-)
Renting your stuff only makes sense if the price isn't ruinous. With Microsoft's cloudy fuckgasm the price is beyond ruinous and well into "obscene." Microsoft's NSA-friendly cloudjsm only looks even remotely attractive if you
A) believe that upgrading every generation is a benefti, to which I I say "Vista, Exchange 2007, Ribbon Bar, Windows 8, VS colour-free edition" and so on and so forth.
B) You upgrade every two years. To which I say "I still have an active installed base of hardware and software over 10 years old and it works just fine, thank you."
Microsoft is trying to use their cloudspooge as a means of making people more, more often. Oddly enough, people want to pay less and pay that less often. SMBs believe a reasonable refresh cycle is 6 years, maybe 10. Microsoft believes it should be every 2 years, maybe 1. And retraining everyone to deal with their lobotomised-monkey interfaces is the problem of the client. (You will adapt to service Microsoft!)
In what universe is that remotely equivalent to leasing a car?
You lease cars you get a choice. You can even hang on to the existing units by extending the lease. You don't get these kinds of choices in the GOOGLE/MICROSOFT/ETC STOP MOVING MY FUCKING BUTTONS cloudy touch-enabled testicle-waving kinect-based future.</ragequit>
NSA, Client/Server, and Big Iron
First, does anyone really want to give the NSA easy access to every scrap of information they own as they move data back and forth from the virtual client to the actual servers?
Which brings up another point, children. This is nothing new (other than the use of the term "virtual" and any non meteorological references to "cloud". Client/server computing (which invented the phrase "thin client" was going to free us from the massive PC investment decades ago. Before that, your virtual machine was a 3270 interacting with an IBM mainframe which did all the heavy lifting. Neither killed the x86 Windows boxes that fill the world's cubes.
VDI Pis a poor solution
One of its many failings is that it needs a lot if disk to make the user experience acceptable. This very expensive disk has made disk manufacturers excited. It really hasn't helped with the financials of VDI vs fat desktop.
VDI - a solution waiting for a problem.
Just wait until Apple starts shipping their new desktops. Then your story will be "the rebirth of the pc."
- Nokia: Read our Maps, Samsung – we're HERE for the Gear
- Kaspersky backpedals on 'done nothing wrong, nothing to fear' blather
- Episode 9 BOFH: The current value of our IT ASSets? Minus eleventy-seven...
- Too slow with that iPhone refresh, Apple: Android is GOBBLING up US mobile market
- NASA to reformat Opportunity rover's memory from 125 million miles away