PCs

Cloud-based desktops a reality

Cloud computing can also include the desktop experience. IT Jedi Rick Vanover outlines how one company is providing a desktop cloud experience that is fundamentally different than other cloud offerings.

Whenever I mention cloud technologies, there are a few familiar retorts to the technologies. This is to be expected, as rarely can we advocate a technology that appeals to the broad TechRepublic family. Recently, I’ve come across a cloud-based virtual desktop offering that will make you think twice before shrugging off another cloud solution.

Providing the client computing environment is the fundamental obligation of an infrastructure group. Further, the client computing environment is a very sensitive nerve that must be delivered well. Considering that current desktop and notebook systems are quite powerful, this is a formidable task.

Recently, I came across iland’s Workforce Cloud offering, and it changed my perspective. One of the obstacles with cloud offerings is the associated requirement to bolster connectivity to reduce latency to server resources that may be in the cloud. The iland approach is to put both the server and the desktop in the cloud. The connectivity between these two infrastructure zones can be Gigabit Ethernet or 10 Gigabit Ethernet. Between the two infrastructure zones, the latency is reduced with the rich connection. The experience from the virtual desktop hosted in the Workforce Cloud is delivered as a presentation technology to any endpoint.

Like any virtual desktop solution, not all situations are applicable for cloud-based presentation of a virtual desktop. The key examples are multimedia rich or graphic design situations where local video processing makes more sense for a high-quality experience.

The iland approach is to be different than other cloud offerings. Instead of being API-driven for its cloud solution, it is more of a co-located infrastructure solution. This extends to physical gear as well. One example of the increased level of engagement would be a mail-filtering appliance for an Exchange server. With iland’s offering, you can put one of the mail-filtering appliances in both your own data center and the iland datacenter to make your disaster recovery or load-balanced cloud-hosted footprint carry the load. This allows hosted desktops to get the same end-to-end experience. The connectivity is driven by customer requirements for VPN for both client and site connections, frequently using SSL-based tunnels that iland has been offering for years.

Do you see this fundamental difference in hosted infrastructure? I am compelled to think that provided infrastructure is more attractive when an end-to-end solution can be architected in the cloud. Let me know if you want to see more of this solution, including a demo and screenshot. Also share your comments below on the thought of a cloud-hosted desktop? Bring your own PC (BYOPC), anyone?

About

Rick Vanover is a software strategy specialist for Veeam Software, based in Columbus, Ohio. Rick has years of IT experience and focuses on virtualization, Windows-based server administration, and system hardware.

40 comments
spydera
spydera

I fail to see anything new here since the days when we had THIN CLIENTS. How is this any different? Also, by including the client in the cloud, what is the technology breakthrough? Everyone and their grandma knows that on a LAN with a Gigabit card, you can get the required speed for anything. So we continue to be fooled by stuff which is just old hat in a new box and costs more. While I can understand the various pros and cons, what happens when the network goes down for whatever reason? Everyone takes the day off I suppose. Not that I am against days off :)

Deadly Ernest
Deadly Ernest

in 1970, just we change who owns the big computer it sits on and now we don't know where it sits. A dumb terminal is a dumb terminal, regardless of where the actual computer doing the processing is sited. And this solution still doesn't give any real protection against all the security and protection issues raised in the past discussions on this subject. Oh, you bets also ask all those people who use T-mobile and Sidekick if the Cloud is still a good way to go with data storage and system usage.

adventurer
adventurer

Yes, this kind of offering is not new - numerous companies have been offering Hosted Desktops (aka Cloud Based Desktops) as a service for many, many years. However, they have typically been very expensive (starting at $100/user/mo.) thus making the ROI challenging for such services and limiting the market to specific industries or needs. What has changed recently in the market is that VMWare has done a tremendous job with creating buzz for VDI (Virtual Desktop Infrastructure) for their View product. The buzz has caused businesses to take a fresh look at Hosted Desktops as a service. Because of this new interest and market potential, we launched NextDesktop.com to offer Hosted Dekstops at an affordable price - $39.99/user/mo. to make Cloud Desktops financially compelling for a wider range of businesses and scenarios. But, keep in mind that not all desktop services are truly "Cloud" offerings. Key attributes of a "Cloud" service are: on-demand, scalable, and self-service. The older generation of offerings require manual provisioning which means that it takes hours or days to get a new user or application setup. NextDesktop is true Cloud solution: this means that customers sign up online, have desktops live in seconds and add applications like MS Project, MS Visio, MS Exchange Server, MS SharePoint, MS Dynamics CRM, and even BlackBerry Enterprise Server through an automated provisioning system. This is all done via a self-service web-based control panel for full control. Customers are in control 24x7 at www.NextDesktop.com - check out the prices and features. Cheers, Ravi Agarwal NextDesktop.com

mikifinaz1
mikifinaz1

I constantly marvel at people who want to go back to being lead around by the short hairs. It is like prisoners who refuse to leave their cells. I didn't like so system administrator telling we what to do in the bad only mainframe days and jumped at the chance to have my own PC even though it meant that I had to do all the work. I guess it is like freedom, most people want to do what they want to do, but don't want the responsibility that comes with it.

bsorenson
bsorenson

Our company, IVDesk, (www.IVDesk.com) offers this solution and has for over 8 years. We host all of a customer's applications, data, include SPLA MS Office licenses and charge a monthly fee per user. We include 24 x 7 helpdesk for all users and continue to have dramatic success adding customers. We've used a number of different technologies but to the end customer the most important things are they get access from anywhere to a hosted desktop with all their applications and data and that everything's taken care of. Not to sound like a commercial but this is out there now and working great for companies of 2 - 200 users without issues. Yup, full motion video doesn't cut it but not that many companies are creating that. It's been here for a long time and we love the "Cloud Computing" focus! Thanks, Bill Sorenson CEO - IVDesk

mredgar2005
mredgar2005

It's the best possible answer for implementing DR, IF done correctly. T-Mobile's approach was totally incorrect, obviously, which caused their data loss. This technology would be ideal for law firms, who need basic desktop and not too much computing power, no graphics, etc. A demo would be pretty interesting, but I still think VM's vSphere is going to end up on top when all is said and done.

jmarkovic32
jmarkovic32

The consumer space is really where Cloud Desktops could take off, in my opinion. The enterprise environment for the meantime, seems to be where cloud desktops will exist in a limited fashion until the mulimedia experience is improved dramatically. I'm really interested in the consumer space, however where this can really take off with people who are turned off by the total cost of ownership of a full-fledged desktop system. With the introduction of inexpensive netbooks and possibly thin clients which are made more affordable with the cell-phone business model, I think cloud desktops could be huge. However in the consumer space, this is contingent on the increase in multimedia performance, namely screen refresh performance for gaming and streaming video.

jck
jck

This is not a new "approach". Hasn't anyone ever heard of terminal services or remote desktop? Remote client access with access to data inside the server farm which are co-located? Besides, then you have to buy a "machine" at their site, storage, then get basically a dumb terminal at your location. More cost...more bandwidth needed to pass graphics back and forth, rather than just data in XML packets...etc. Again, cloud computing fails. Not cost-effective for the majority, not efficient in the slightest, and not 99.9% dependable like most in-house solutions. It's a niche market, at best

reisen55
reisen55

The frightening aspect on this one is that cloud server computer merely deletes a server or virtual server farm with loss of data and time. Now if we get into cloud desktops what happens if, as it does, the network goes down. Instead of at least partial functionality on a limited stand alone basis as is traditional IT when the network drops off (router error, port issues, etc), you can potentially lose the whole desktop? At what cost does this mini disaster scenario become cost-effective????

tony
tony

While this approach makes a lot of sense, is it fundamentally different than a farm of MS Term Server / Citrix servers? I assume that you "lease" the equipment / access instead of own it. Tony D

darcdclub
darcdclub

I agree with you, also I would like to view a demo or at less some screenshots. Thanks

mrobas
mrobas

great article. You might want to check www.flipit.si It's a cloud based solution extended with a hardware service offer also (for thin clients, printers, voip, ...) And it does incorporate both servers and desktops in a cloud. A true service for a customer. BR,Mitja

b4real
b4real

I may check that one out as well.

jmarkovic32
jmarkovic32

Who would go with a cloud services provider who didn't have co-lo facilities and other redundancies?

jmarkovic32
jmarkovic32

...is the long term cost savings on hardware. This is significant for places like large hospitals.

b4real
b4real

Sorry for your difficulty interpreting the piece. Mainly, I'm tossing out the idea and seeing what people think about hosted desktops.

b4real
b4real

The wire is the most important piece at that point.

jmarkovic32
jmarkovic32

Every solution has a pro and a con. I think the long time savings outweigh any potential problems though. There is a lot on the local end that you can do to prevent this problem and there's alot the service provider can do on their end. Many cloud providers (should) have redundancies built in at almost every layer from redundant hardware to colocation facilities. On the local end, dual routers and dual ISPs can limit that risk to near zero. The cost of that is almost negligible in comparison to a self-hosted solution when you look at the overal TCO.

mark
mark

When you are managing thousands of PCs the trade off can become correct. Take H&R block, they provide tax services, every year each computer is wiped once and set to a factory update. Consider the same service with dummy terminals. Now you no longer need to upgrade hardware at each prem, instead you have dummy terminals with perhaps 1-2 full PCs per location instead of the 8-10 currently allocated. Employees at this site would not need to save anything to the machine or cloud. It is just interfacing with a program and submitting the results to the IRS and logging all of that. All that data is already centrally stored and backed up anyhow. This could prove workable for this type of environment. The larger problem is going to be bandwidth, latency between cloud and users, and internet outages. Currently a H&R block can function in a limited fashion when the internet is down. Should it be down for this system then the whole works is done. What the cloud really needs is a hybrid system that has a local server which connects to the cloud to sync databases and perform updates. This way bandwidth utilization is minimized and end user experience is maximized. This would also be critically important if the site is using any voip solution in a scenario where bandwidth is expensive.

spydera
spydera

Here are the main differences: 1. More complex terminology to define the same stuff you have been doing for ages 2. Higher Costs of everything 3. Absolutely no guarantees of data security 4. Absolute certainty that your stuff (data) is out there waiting to be: ---a. Hacked ---b. Can land in some blackhole of no recovery ---c. Takes forever to access 5. And you probably don't own anything, just lease it so its a recurring revenue model instead of a one-time cost. In item 5, I see endless opportunities to milk the poor user for all kinds of stuff. Anyone care to add a list you can be milked for? How about data recovery charges!! Extra Backup charges .....

b4real
b4real

The underlying technology for iland is VMware-based.

b4real
b4real

I will see about getting those going. Thanks

LocoLobo
LocoLobo

Is there an english version of the site?

jck
jck

But what is the cost savings when the medical systems go down, and patients lives are at risk because someone backhoed through a fiber conduit? Critical systems should never be handed off, no matter what the cost. That's poor risk mitigation, and poor management does that. Assure your necessary assets, and strive to get the most optional ones at the lowest cost. The day you start letting critical systems' dependability lapse is when it borders on negligence. And in something like healthcare, negligence goes hand-in-hand with lawsuit.

blarman
blarman

As far as I can see, you're not really reducing the hardware footprint - merely the support costs. You've still got to have a computer to use to access the apps regardless of their location. And really all you are doing is trading support costs for networking infrastructure. I'm sure that there are some businesses that will be able to utilize this type of infrastructure to their benefit, but it is going to be a niche market.

darpoke
darpoke

following you and I'm neither a database administrator nor do I maintain a server farm. We have a single Xserve here that acts as a simple fileserver. I maintain a Vista VM on my personal work machine that's only used to test our output from a PC environment. Personally I think the cloud has a lot to offer - always have - and that we're only recently approaching a level of infrastructure where it can even be contemplated. Once network bandwidth and stability reaches a critical mass people will be all over this technology. How quickly we forget the days of payphones. Ask someone back then whether they'd expect to be making calls from something smaller than their wallet and they'd have looked at you like you were mad. Tell them they'd have full internet access from the same device - even in the early days of people know what the internet was - and they'd have had you sectioned. It still astonishes me, the lack of foresight, and the level of closedmindedness that debate over the viability of the cloud provokes. In ours of all industries we should understand *and value* the pace of change. And progress!

wabernat
wabernat

when we had a 3745 in the closet using an sdlc ciruit back to the mainframe where all the email and application stuff was kept. Every one of those IBM terminals was diskless except some of 3276s had a 8 inch floppy drive cause the controller and terminal were in the same box. Or was that a big VAX vms cluster in the data center with vt100 terminals running all-in-one or some FMS linked to Datatrieve database app... The more things change the more they stay the same.

darpoke
darpoke

- should the need arise, the service provision can be scaled up in little to no time. Contrast this with the time needed to boost local resources, and the question of what to do with them if the increased processing bandwidth was temporary. This approach has definite issues that should give one pause, but (i) so does every approach, (ii) many approaches have similar concerns but are routinely done without apprehension, and (iii) this approach brings several benefits to the table that others lack. I understand the attitude that people take with mission-critical resources - protect at all costs - as without this our industry would never have become the trusted facilitator of most business that it is today. But ours is an industry where paradigms can be turned on their head faster than almost any other - we should be proud of that and try to avoid the automatic closing of minds that the question of the cloud always provokes. A closed mind in this business is rarely the optimal approach. Why start adopting it now?

jmarkovic32
jmarkovic32

I'm not addressing this to you specifically, but detractors of cloud desktops pretend as if these concerns are new when they are not. Alot of companies run Citrix or TS to publish apps to remote sites. What if there is a WAN issue in that situation? I'm completing a TS RemoteApp application rollout and I had the same concerns beforehand. However, we looked at the facts: WAN outages were a rare occurance and even when they happened it would be solved in a few minutes. Most of our issues were related to server issues/DR concerns at remote sites. We have limited IT staff (just me) and so the service levels were horrible. It would take half a work day (at the least) for me to fix server issues for some sites. So my decision was easy. In the event that there is a WAN outage, our developer created a program that downloads some data nightly to pre-designated computers at each location. With it, the users can access a local streamlined version of the app while we work out the WAN issues. I don't see why this concept wouldn't work in most situations. There could be one fat client machine at the location where a local copy of the data can sit.

b4real
b4real

The ALP (Application Link Protocol) is a very mature presentation technology, good call out. This is a VMware-based solution in particular from iland, however.

billyg
billyg

I'm surprised to see no mention of SunRay in this discussion. It is an interesting model in that it addresses a number of motivations behind cloud computing, but places control/security in private hands. Also worth noting, we tend to stop scrutinizing power consumption outside the data center, but consider that a SunRay dissipates 4 watts of power. Multiply that reduction factor by the number of desktops, laptops, towers in your organization... If rich media experiences don't include a lot of high speed 3D rendering, then SunRay covers that pretty well, too. That all said, my point is not to advertise for Sun/Oracle. This comment is only to emphasize a particular thin client model that might be thought of as a privately managed cloud. It so happens that this is a very proprietary model with only one vendor. It is the model that I think has merit.

jck
jck

But which is more likely to be hazarded? The wire running above the acoustic tile in your ceiling or down the walls? Or, the fiber-optic line running down the side of I-75 that feeds both Level 3 and AT&T's backbones and has massive road construction going on near it? I'd depend on a Cat5e/6 inside my building more than a fiber line 100s of miles long at a datacenter miles away. As I'd said above: Cloud is going to end up a niche market. Places like confidential data holders (government, banks, etc) can't use it because of security, and people with large scale data with mass users won't be able to because the wire won't handle remote desktop access for 5,000 end users. Besides that, what's faster in determining if your service is down on a server? You going to the console, or having to call administration support and waiting on hold 15 seconds to 10 minutes based on your priority and severity and call volume? If you want maximum (not total) dependability, you're better having things in-house. That's my point.

b4real
b4real

Anytime ANYTHING is centralized, there is a risk of remote sites going down. Arguably, we always depend on the wire.

jck
jck

So what's more expensive? Running a hospital with an on-site data center or Running a hospital with a seperate data center that requires as much (if not more) in hardware costs (servers aren't cheap) as well as 1 (or more) data trunk lines? Plus, you have disconnected staff. Efficiency drops. You can't multitask your staff as well when operating two sites, or you have to increase payroll costs. There goes your cost-effectiveness out the window. The likelihood of a mission critical interruption happening on-site is less than one happening to interrupt service for a remotely-functionalized architecture. That's why I mentioned risk mitigation. If a natural disaster hits your single facility in a hospital, you have a LOT more to worry about than your server going down. But if that natural disaster hits your offsite systems, then you have to not only worry about either shuttling staff there to do it, or getting in contact with those onsite to get things back online somehow to function (which in a natural disaster is pretty much impossible). And not being able to accurately access medical charts and records is problematic in the proper treatment of your patients. Added comment: "Cloud" isn't new either. I've been connecting clients to remote datasets for 15 years. Just a new term for an new way of doing the same old thing.

jmarkovic32
jmarkovic32

A major hospital in my area has a datacenter offsite that it owns. The hospital connects to critical apps via Citrix. If communication lines are cut, they lose access to systems. That's why they likely have another WAN link from a second ISP as a backup and the datacenter probably has one as well. We can think of any doomsday situation we want. Some of us get paid to do so. However, it doesn't change the fact that this isn't a cloud or no cloud option because some hospitals/enterprises already access critical systems remotely. I can think up a million doomsday situations for hosting data onsite. What if there's a natural disaster? What if there's a fire? Security and privacy concerns may be legitimate NEW concerns when it comes to cloud computing, but system availability is not. Companies like mine (albeit small) have the same concerns and we host our own environments.

jmarkovic32
jmarkovic32

Thin clients are significantly less expensive compared to full-featured desktops. On top of that, support costs are almost nil. Granted, the initial investment may be large (you still have to buy the hardware and pay for the service), but by the next refresh cycle you should double or triple your ROI once you consider all of the cost savings (power, hardware, hardware support).

LocoLobo
LocoLobo

At least I learned something new today. I didn't know about translate.google.