In my tenure at Intel, I have had the pleasure of walking into major companies, educational institutes, non-profits, and government agencies to talk technology with many great people. “How green is this solution” is a topic on many minds lately – no matter which topic of discussion. Being an engineer by trade and scientist by education, I will typically dive into the details of around each component’s power consumption and the discussion ends with some simple math multiplying a number of units by their thermal numbers. However, there is so much more to the overall impact, and as I walk in and out of these locations, I am always amazed at the number of larger issues with much larger impacts that are unresolved or overlooked. Reading the book “Living Like Ed: a Guide to the Eco-Friendly Life” by Ed Begley Jr., inspired me to approach some of these topics, and to similarly classify items by their degree of difficulty to implement – easy changes, not-so-big changes, and big changes. Additionally, I will belooking at the overall impact that compute model choices can affect. However, I will leave the topics beyond the realm of compute models to experts such as Ed Begley Jr.
Corporate Recycling, it can be an easy change:
Before I dive into any of these subjects, recycling should come as an essential component to every one of these solutions – it should become part of your culture, the stockholders will most likely appreciate the frugality. If your purchasing new equipment, you need to be thinking about what you can do with the old equipment, sometimes the answer is to donate the equipment to charities, sometimes it needs to be disposed of, but rarely does that require it fill a landfill. As an example, everywhere that Intel operates, more than 70% of all waste is recycled. I am not suggesting you need to achieve this overnight, Intel has been working on this since 1971…it is a gradual process. Start by looking at what the biggest waste items are from your company and get creative – is it finding a use for all those coffee grounds, finding ways to reuse packaging material when shipping your products, or simply implementing recycle bins and growing employee awareness.
Here is a great video which highlights how Intel practices corporate recycling:
Pure power consumption items:
Monitors
I have yet to find a location that I have visited where I cannot find that amongst the rows of office workers, several are still using CRT monitors – and many times they are not even the energy efficient CRTs. Simply moving these users from CRTs to LCD can have a profound impact on power consumption. Consider a typical 17” CRT will consume around 80 watts, and a 15” LCD is around 25 watts (these have similar viewing areas). For any user working behind one of these outdated CRT monitors, we need not discuss any other aspect of power savings at their desk until this is fixed, no compute model savings are looking to give you 55 watts back with such a simple solution. Added to this are well-known benefits around increased worker productivity when moving from CRT to LCD due to eyestrain reduction, glare, distortions, flicker, and visual search time improvements. As far as I can see it, switching out these monitors is an easy change, it is in the same vein as moving from incandescent to compact fluorescents light bulbs. Their are even HVAC efficiency changes when these changes happen on a large enough scale (less heat put off by the monitor equals less cooling needed from the HVAC – and in winter the heating produced by your HVAC system I am going to assume is more efficient than the heat being produced by that CRT). http://ergo.human.cornell.edu/PUB/LCD_vs_CRT_AH.pdf
However, a big change item enabled by the CRT to LCD upgrade comes in the realm of building design. The distance an employee sits from an LCD is the same as the CRT, however the space needed behind the LCD is far less than that of a CRT – LCD monitors even have direct wall mount options. This gives space designers the ability to decrease desk depth and develop creative solutions around ergo designs. This results in more compressed, configurable, and/or productive work environments. The weight reduction on a given office floor can give some relief to building designers as well (average CRT weights 40-45 lbs, and the similar LCD is 6-8 lbs – multiply this by the number of workers in a building, let say 1000, yields a couple tons removed from a single floor). Is their a way to utilize the weight and heat to balance locations that are often constrained already, such as your server room? It’s worth looking into.
Telecommuting
The subject of telecommuting in general causes various reactions – from the employer who has witnessed abuse of the telecommuting freedoms to discussions around increased employee focus and higher output, the reaction and debate on this subject will continue, just as it has around anything from solitaire to YouTube and social network use. Regardless of the outcome of these debates, telecommuting can have an overall world effect on the number of employees on the road to and from the office each day. Being that most people tend to take a transportation method that is far from efficient, this alone can be a net positive impact. I have heard some government employees are encouraged to spend 1 day each week working from home (pending their job allows them to do this) simply to reduce the environmental impact. However, the debate is still out on the efficiencies in power consumption regarding heating and cooling a single residence verses several employees in an office environment, and the infrastructure costs to support more remote verses local employees. The benefits of mobility in your compute model can definitely benefit the environment – at the least, mobility offers the flexibility to consider various work environments (e.g. what would be greener than a person using a laptop outside using solar power?).
Power policies
Often power policies have taken energy efficient configurations and pushed them another 20% beyond what is already seen as good. When applied down the wire over manageability interfaces such as can be done on an Intel® vPro™ Technology enabled client, they are easy to deploy and quick to update when needed. You can decide to simply turn off the unused computer, wake it up and update it when needed, and then return it to the low power states. On the other hand, you can utilize the processing power of that vacant machine to run a distributed compute environment using an IDE redirection operation, further reducing the loads on your data center and switching the watt per calculation onto inexpensive devices. Isn’t this the whole argument that drove RAID technology – the I in RAID stands for inexpensive, we took inexpensive drives and made redundant copies, much like we can do with the relatively inexpensive computations of these vacant client machines.
Data Center Consolidation via Virtualization, DC Racks, and other things inside those glass rooms…
If no one has looked into this area yet at your organization, it may be time to visit your local datacenter and see what is going on. Chances are the ladies and gentlemen running your datacenter are already plugged into these topics, as they are probably spending a large budget every year just to keep those servers that you never see humming along, money well spent on temperature and climate control, money spent on power consumption, etc. Lets face it, unless you are someone who understands the difference between 1U, 2U, and 4U and knows what happens when the halon system is engaged, then you should get your teams that do know about such things looking at The Server Room. There are many fantastic advances in the last few years that can drastically decrease the power consumption, reduce the cooling needs, and increase the manageability and reliability of your server room.
Compute Model Debates:
The reason I call this section “debate” is that various individuals, corporations, analysts, and product vendors spend much time debating about which of these is greener. The argument usually stays within the simple math as I described before, but the real answer I believe should extend into the larger, holistic, picture. All of these solutions fit into the big changes category, as they require establishments to modify the way they operate, often involving the acquisition of new equipment and software, and typically requiring end-users to receive some training to function productively in these environments. For more information on these compute models you should read the presentation at: Compute Models Explained
Fixed Location: Terminal Services, Virtual Hosted Desktops, Blade PCs, and Web-Based Apps
I am grouping all of the compute models that move large amounts of computation from clients and place them on server(s), as they all have a very similar green impact with slight nuances related to each one. Often this group of computing wins out quickly with simple math:
Current: 20 client computers running at X watts + server running at Y watts
Thin-client: 20 thin-client computers running at almost no watts + server running at Y watts
This always looks great, and why not, you are reducing the wattage on the item that is being multiplied. Too good to be true? This does not account for several key items with these calculations. The scenario does not look at how many clients the one server handled before and after the switch. The change pushes more computing demands on the server, and reduces the demands on the clients. In the current scenario, the server may have been handling very limited calculations, and could have stretched to thousands of clients. Anyone who has priced servers and supported them knows that the cost per calculation that you pay on a server is far greater than the cost per calculation on a client. Servers require redundancy, are often located in raised floor climate and temperature controlled environments, typically are allowed to operate at up to 50% capacity before more are added to the mix, are supported by disk arrays which are also climate controlled and redundant, have built in fans with redundant fans, built in power supplies with redundant power supplies… All of this is to make sure that you, the end user, never experiences downtime. On the other hand, your desktop or mobile client is built with the end user in mind, it can often handle limited shocks, a wide range of temperatures, humidity, and electro-magnetic interference. However, it does not have dedicated employees supporting it as the server does, and does not require a special room. I have yet to hear a client discussion where we talk about five or six 9’s of uptime – client computers simply reboot much more often.
The real equation should read something like the following:
Current: 2000 client computers running at X watts + server running at Y watts
+ server room HVAC
Thin-client: 2000 thin-client computers running at almost no watts + 200 servers at Y watts
+ server room HVAC expanded to support 199 more servers
Is this a net wash, increase, or reduction – that depends on how constrained your datacenter already is, what part of the world you are located in (do you have some glaciers nearby), and several other factors. I am not saying it is not always a net reduction of power, but the equations used are often over simplified, they have to take a holistically approach to each environment to determine the true merits.
Off Network Options: Distributed, Rich Client, Virtual Containers, Application Virtualization and Streaming
The second group I am going to split my debate into is the group that supports mobility and retains computation on clients. Many of these models support moving computations between the server and client as dictated by policies and system capabilities. However, in general each of these models enables the server component to scale to a much larger number of clients, and when reaching server capacity can use policies to turn up client compute loads. The same calculations apply in these environments where you include HVAC costs for the increased server demands. However, the numbers of servers increased in these scenarios are much smaller. Don’t just take my word for it, follow this linkto a study done by Fraunhofer Institute
Conclusion:
No step is too small…changing behaviors, deciding which solutions are right for you, and reaping the benefits of growing greener is a gradual process, one for which we all should strive. With each option we need to be looking at what the larger impacts are – what does it mean to productivity, security, manageability, and is now the right time to gain adoption for this change?
Emerging Compute Model Forum
Let me start by explaining where many of us are now. We typically live in a world where we boot a computer to a rich operating system that has many features we may or may not use, then we install applications off CD/DVD, downloading installers over the internet, or have it pushed as a local install over the corporate network. We all run local virus scanners, firewalls, and patch/update everything often - less we fall behind and become vulnerable. Each of these software programs we use have been tested to work with our operating system (or we hope), but very few of them are tested to work with other applications, and some are just not compatible with each other (they didn't make it out of kindergarten with the "plays well with others" moniker). Many programs are installed on a ton of computers, with much of the data being the same across those computers, but being that content belongs to the person next to you, it is redundant but not accessible (across a given large group of people the amount of duplicate data is enormous...larger than having copies of the US library of congress in digital format). Some people have started moving away from this model, but often come up with solutions that are either too awkward to become mainstream, or too limited to become useful.
Next, the path to the future... With several compute model choices, people have started using the modern day compute and network resources to revisit solutions that had limited success in the past (I say limited as none of them won out over the model described above, many were very successful in specific environments). To help with the large amount of redundant data, people moved the data to server rooms. To deal with application conflicts people gave each of these programs their own virtual sandbox to play in (now they don't have to play well with others...they get their own sandbox instead). To deal with patches and updates, people developed utilities to maintain compliance with a few button clicks (and several scripts, settings, and close monitoring). And the list goes on...
The future... Now I will put on my rose colored glasses and look at where things are going...in other words, I believe they are taking a turn for the better. Going back to the discussion that was had during the BriForum class, the basic architecture was a "dial-tone OS" with virtual containers that can be streamed and executed locally or presented over the network. The term dial-tone OS was new to me, and as I believe Ron Oglesby described it, the operating system would give a basic level of functionality similar to when you pick up a phone and hear the dial tone. We all have grown to expect a dial tone when you pick up the receiver, and if there is a pause or delay we are very confused as we have grown a very high level of expectation for the quality of service on this device (not talking about coverage areas here - just the basic features). With a dial-tone OS, the client device would quickly respond with some basic features - a GUI (graphical user interface)/window manager, a scheduler, I/O mapper, device drivers, and a virtual machine manager (I may be missing a few OS fundamentals, but the idea here is a truly minimal/microkernel type OS that has a high level of reliability). All application that execute in the environment would work in their own virtual sandbox, which may contain an entire OS emulation, or simply the basics to execute - or in other words virtual containers. These applications would interact with the GUI via the window manager, and negotiate the layout within the systems capabilities. The Virtual Container would execute either locally, on a server, or in the network/cloud based upon the negotiated policies and client device capabilities. For containers executing locally, differences of the container would be archived and ready for use on other machines or as backup (depending on connectivity, etc).
The key here is an environment that from the base up is built with device capabilities in mind - if you're executing a spreadsheet calculation and your device is going to take days to calculate it, have another location process that for you. If you're using the same data as everyone else, make one image of it in the community of users, and everyone works from that image - when the image is upgraded, everyone migrates over time. If the device has Intel vPro capabilities, the virtual containers and dial-ton OS can take advantage of the energy-efficient performance, manageability, and security features. If the device is ATOM based, then a whole new set of features are exposed. Etc... (I had to add in my own Intel fanboy comments, but comments I really believe in).
The road to get from here to there involves a ton of non-trivial solutions, and I believe the good news is that many of the solutions are being thought about by some great minds - however I am sure there are some new and exciting "change the world" ideas left to solve...
The future looks both responsive and reliable, and environment where we are not encumbered by the limitations of our environment, but simply a click away from doing our next task.
-Jason Davidson
Facebook, Twitter
Lately I have been really into the future of virtualization. The concepts that I have been learning in school really didn't sink in until I was thrown into it. It's funny how that works. I am not saying getting a higher education that you wouldn't learn anything. I am not saying that, you learn a lot. But what I am saying is that some people's passion goes beyond than what you learn from books. I can go to a French class everyday and learn the language. If I go to France then I can learn the culture and the language. What I am talking about here is Immersive Studies in Virtualization!
One trend that has become clear to me with all the cool hardware I get to see in my internship at Intel is that the hardware gets stronger while the size and power requirements gets small - and this is not going away. But one thing we have to realize, there is always going to be the equipment no matter what, and that equipment is going to have more and more features for us to pound on. Virtualization has been my new love, and not just for server consolidation, but application and desktop virtualization are the next killer ideas. The concept really sunk in after a few talks with Jason Davidson (my guide through the galaxy, you are my virtualization 42!) and also during the BriForum (I am still on a high from that one!!).
In my other blog series on the vPro Expert Center, I am on a journey of learning vPro (links), and feel like intellectually I am on a roller coaster ride of knowledge. Now that I am on this virtualization roller coaster I have to wonder how wild and crazy this ride will be - am I going to get off it alive? Application and desktop virtualization I believe will soon start to take over our lives.
As this old Commodore 64 advertisement below portrayed...25 years ago, people wanted to be the "movers of this world" with the power at their fingertips. This was typically used to simply play games and create a few documents. Now, 25 years later, the world has changed - we have mobility, we still have fun, but we stay connected. Yet, we are still on this ride of looking for the next faster, smaller, cute, reliable, and fun, device - it almost has to be our non-emotional twin.
The computer reflects who we are as a person. The applications that is on our systems fit to how we like them. Is it going to come to a point where we can just think of what we want and the computer will know automatically what it is? (I guess that's what Google is for but you still have to type it out) it's to the point where we won't need to have the computer in front of us. We will be able to talk into a "Bluetooth" type deal and all the information will come up on a mobilized screen in front of us. (Huh, okay nobody take that idea I am going to go out and patent it right now!!)
But now let's look at how application and desktop virtualization will deliver some of this for us... We moved from the floppy disk install to the cd/dvd install, and soon we will not be installing at all! A great reference is the video about St. Agnes Academy. (Check it out on Jason Davidson's area St. Agnes Prep School use Emerging Compute Models with Video.) Will the operating system become less important, and provide just the basics we need to launch any application in a virtual environment - an environment that we can have upgraded & managed with ease. Will future computer users never install application - but simply click the icon to launch them and boom you have it? How easy is streaming going to make things?
Add to this the vPro features and I can see a day when IT doesn't have to physically be in front of your computer to diagnose or fix it, and when it is broken you can migrate to a new one with some simple streaming...
Pretend for a moment that you had the opportunity to come up with anything in the world, anything at all. You had all the equipment and can make anything. What would be your item that you would make to revolutionize this world?
Now the scenario - while on this trip one of the PC's who is up to date with virus protection and patches developed a virus, and as much as I would like to spend the time looking into how the virus got there - doing this over the phone would not be feasible. Therefore, I did what any modern day geek would do - I VPN'd into my home from my hotel, I took control of the computer over a remote desktop session and started fixing. I found the virus engrained into the system, and to keep my home running until I return, I set the machine to boot to the network instead of the local hard disk using IDE-R (a feature in vPro). Then I rebooted the machine and it booted Ubuntu Linux over my network, and the files that my family uses are accessible over the file shares.
Problem patched - until I return home. Keeping my fingers crossed...
-Jason
p.s. If you have any questions on how to configure your house this way - fire away.
If I had to identify the state of mind of the many IT managers we meet regarding emerging compute models, that's it. Technology is changing quickly. Buzzwords abound. IT shops are experimenting with just about every model and every combination thereof. Many have expressed frustration that a clear path to the future has not yet emerged, but all eagerly desire relief from management, security and TCO challenges promised by vendors hawking each solution.
Another thing we frequently observe is that very few IT shops are aware of all the compute model options available to them. They all understand classic terminal services and rich, distributed clients - the two ends of the spectrum. Awareness of the choices in between is far from universal, much less a comprehensive understanding.
So, I'll attempt to add value to the world by laying out the landscape as we see it, and let you know where Intel stands.
A Simple Taxonomy
When I talk to customers, I parse the compute model choice on three vectors:
1. Where does the application execute?
2. Where is the data stored?
3. Can it function off the network (mobility)?
Using those criteria, the major compute models we track split out like this
I'll admit to a couple flaws in this picture. I change definitions for Blade PCs depending if they are used in 1-user-per-blade mode, or many-users-per-blade. In the 1:1 mode, it is just a repartitioned PC, so I classify them as "client-side". In 1:Many mode, the blades are more like servers. Web-based apps are also not pictured. I think of server-based web apps are similar to terminal services. Web apps that rely on client-side execution resources, such as Java, Flash, Ajax and Silverlight, behave like streamed applications.
Devices versus Compute Models
"Thin client" is a loaded term. To some, it's a software architecture. Others see it as a specific device - a terminal. I'm always careful to separate conversations about which compute model is appropriate for a situation from the client device. A desktop or laptop PC can act as a terminal or "thin client" when using server-side applications. My recommendation is first determine the compute models a user will need, and then figure out the right device.
The Ups and Downs of the Models
The list of pros and cons for each model is long, and many authoritative sources and pundit sites are available to debate them. I'll provide a quick list of what we see as the major benefits and biggest limitations of each.
Terminal Services
Virtual Hosted Desktop
OS + Application Streaming
Application Streaming (aka Application Virtualization)
The Intel Point of View
After laying out these options and the pros & cons, customers inevitably want to know where Intel comes down on it all. Our view is this:
Let the Games Begin
We are living in interesting times. Evolution of computing models is happening in real-time, and there is no shortage of opinions and viewpoints. I hope this blog has been informative. Even if it hasn't, I hope it is a launching pad for a spirited (yet respectful) debate here. Let the games begin!
ECMF Blog
12 Posts tagged with the point_of_view tag
A few weeks back at the BriForum I attended a session called The Future of Client Computing, where the audience participated in an open discussion around where client computing is headed. It was amazing to see a group of very smart people come to a single consensus...with various interpretations of that consensus I am sure. Being that I am back from the show, and back from vacation, I wanted to take a few minutes to recap what my interpretation of the future...
Let me start by explaining where many of us are now. We typically live in a world where we boot a computer to a rich operating system that has many features we may or may not use, then we install applications off CD/DVD, downloading installers over the internet, or have it pushed as a local install over the corporate network. We all run local virus scanners, firewalls, and patch/update everything often - less we fall behind and become vulnerable. Each of these software programs we use have been tested to work with our operating system (or we hope), but very few of them are tested to work with other applications, and some are just not compatible with each other (they didn't make it out of kindergarten with the "plays well with others" moniker). Many programs are installed on a ton of computers, with much of the data being the same across those computers, but being that content belongs to the person next to you, it is redundant but not accessible (across a given large group of people the amount of duplicate data is enormous...larger than having copies of the US library of congress in digital format). Some people have started moving away from this model, but often come up with solutions that are either too awkward to become mainstream, or too limited to become useful.
Next, the path to the future... With several compute model choices, people have started using the modern day compute and network resources to revisit solutions that had limited success in the past (I say limited as none of them won out over the model described above, many were very successful in specific environments). To help with the large amount of redundant data, people moved the data to server rooms. To deal with application conflicts people gave each of these programs their own virtual sandbox to play in (now they don't have to play well with others...they get their own sandbox instead). To deal with patches and updates, people developed utilities to maintain compliance with a few button clicks (and several scripts, settings, and close monitoring). And the list goes on...
The future... Now I will put on my rose colored glasses and look at where things are going...in other words, I believe they are taking a turn for the better. Going back to the discussion that was had during the BriForum class, the basic architecture was a "dial-tone OS" with virtual containers that can be streamed and executed locally or presented over the network. The term dial-tone OS was new to me, and as I believe Ron Oglesby described it, the operating system would give a basic level of functionality similar to when you pick up a phone and hear the dial tone. We all have grown to expect a dial tone when you pick up the receiver, and if there is a pause or delay we are very confused as we have grown a very high level of expectation for the quality of service on this device (not talking about coverage areas here - just the basic features). With a dial-tone OS, the client device would quickly respond with some basic features - a GUI (graphical user interface)/window manager, a scheduler, I/O mapper, device drivers, and a virtual machine manager (I may be missing a few OS fundamentals, but the idea here is a truly minimal/microkernel type OS that has a high level of reliability). All application that execute in the environment would work in their own virtual sandbox, which may contain an entire OS emulation, or simply the basics to execute - or in other words virtual containers. These applications would interact with the GUI via the window manager, and negotiate the layout within the systems capabilities. The Virtual Container would execute either locally, on a server, or in the network/cloud based upon the negotiated policies and client device capabilities. For containers executing locally, differences of the container would be archived and ready for use on other machines or as backup (depending on connectivity, etc).
The key here is an environment that from the base up is built with device capabilities in mind - if you're executing a spreadsheet calculation and your device is going to take days to calculate it, have another location process that for you. If you're using the same data as everyone else, make one image of it in the community of users, and everyone works from that image - when the image is upgraded, everyone migrates over time. If the device has Intel vPro capabilities, the virtual containers and dial-ton OS can take advantage of the energy-efficient performance, manageability, and security features. If the device is ATOM based, then a whole new set of features are exposed. Etc... (I had to add in my own Intel fanboy comments, but comments I really believe in).
The road to get from here to there involves a ton of non-trivial solutions, and I believe the good news is that many of the solutions are being thought about by some great minds - however I am sure there are some new and exciting "change the world" ideas left to solve...
The future looks both responsive and reliable, and environment where we are not encumbered by the limitations of our environment, but simply a click away from doing our next task.
-Jason Davidson
Facebook, Twitter
I am sitting here contemplating what does ECMF have to do with me?
Lately I have been really into the future of virtualization. The concepts that I have been learning in school really didn't sink in until I was thrown into it. It's funny how that works. I am not saying getting a higher education that you wouldn't learn anything. I am not saying that, you learn a lot. But what I am saying is that some people's passion goes beyond than what you learn from books. I can go to a French class everyday and learn the language. If I go to France then I can learn the culture and the language. What I am talking about here is Immersive Studies in Virtualization!
One trend that has become clear to me with all the cool hardware I get to see in my internship at Intel is that the hardware gets stronger while the size and power requirements gets small - and this is not going away. But one thing we have to realize, there is always going to be the equipment no matter what, and that equipment is going to have more and more features for us to pound on. Virtualization has been my new love, and not just for server consolidation, but application and desktop virtualization are the next killer ideas. The concept really sunk in after a few talks with Jason Davidson (my guide through the galaxy, you are my virtualization 42!) and also during the BriForum (I am still on a high from that one!!).
In my other blog series on the vPro Expert Center, I am on a journey of learning vPro (links), and feel like intellectually I am on a roller coaster ride of knowledge. Now that I am on this virtualization roller coaster I have to wonder how wild and crazy this ride will be - am I going to get off it alive? Application and desktop virtualization I believe will soon start to take over our lives.
As this old Commodore 64 advertisement below portrayed...25 years ago, people wanted to be the "movers of this world" with the power at their fingertips. This was typically used to simply play games and create a few documents. Now, 25 years later, the world has changed - we have mobility, we still have fun, but we stay connected. Yet, we are still on this ride of looking for the next faster, smaller, cute, reliable, and fun, device - it almost has to be our non-emotional twin.
The computer reflects who we are as a person. The applications that is on our systems fit to how we like them. Is it going to come to a point where we can just think of what we want and the computer will know automatically what it is? (I guess that's what Google is for but you still have to type it out) it's to the point where we won't need to have the computer in front of us. We will be able to talk into a "Bluetooth" type deal and all the information will come up on a mobilized screen in front of us. (Huh, okay nobody take that idea I am going to go out and patent it right now!!)
But now let's look at how application and desktop virtualization will deliver some of this for us... We moved from the floppy disk install to the cd/dvd install, and soon we will not be installing at all! A great reference is the video about St. Agnes Academy. (Check it out on Jason Davidson's area St. Agnes Prep School use Emerging Compute Models with Video.) Will the operating system become less important, and provide just the basics we need to launch any application in a virtual environment - an environment that we can have upgraded & managed with ease. Will future computer users never install application - but simply click the icon to launch them and boom you have it? How easy is streaming going to make things?
Add to this the vPro features and I can see a day when IT doesn't have to physically be in front of your computer to diagnose or fix it, and when it is broken you can migrate to a new one with some simple streaming...
Pretend for a moment that you had the opportunity to come up with anything in the world, anything at all. You had all the equipment and can make anything. What would be your item that you would make to revolutionize this world?
If you are like me, when you travel, computers break at home - and being a computer person, you are the tech support...your house is most likely your personal lab, in a constant state of flux. If not, I salute you. To make matters worse, I am often the one who messed things up before I leave - luckily, my wife patiently waits for me to get into my hotel and work with her to fix it remotely. She already does a great job at tolerating the wires, keyboards, mice, monitors, and various other computer parts in every corner of our house - so having to wait for me to fix these, is a hassle for her I would like to reduce.
Now the scenario - while on this trip one of the PC's who is up to date with virus protection and patches developed a virus, and as much as I would like to spend the time looking into how the virus got there - doing this over the phone would not be feasible. Therefore, I did what any modern day geek would do - I VPN'd into my home from my hotel, I took control of the computer over a remote desktop session and started fixing. I found the virus engrained into the system, and to keep my home running until I return, I set the machine to boot to the network instead of the local hard disk using IDE-R (a feature in vPro). Then I rebooted the machine and it booted Ubuntu Linux over my network, and the files that my family uses are accessible over the file shares.
Problem patched - until I return home. Keeping my fingers crossed...
-Jason
p.s. If you have any questions on how to configure your house this way - fire away.
Video Interview - Ketan Sampat, Marianne Jackson, and Arjun Batra
Posted by Jason A. Davidson Apr 9, 2008
Observations from an Intel Guy: Mike Ferron-Jones
If I had to identify the state of mind of the many IT managers we meet regarding emerging compute models, that's it. Technology is changing quickly. Buzzwords abound. IT shops are experimenting with just about every model and every combination thereof. Many have expressed frustration that a clear path to the future has not yet emerged, but all eagerly desire relief from management, security and TCO challenges promised by vendors hawking each solution.
Another thing we frequently observe is that very few IT shops are aware of all the compute model options available to them. They all understand classic terminal services and rich, distributed clients - the two ends of the spectrum. Awareness of the choices in between is far from universal, much less a comprehensive understanding.
So, I'll attempt to add value to the world by laying out the landscape as we see it, and let you know where Intel stands.
A Simple Taxonomy
When I talk to customers, I parse the compute model choice on three vectors:
1. Where does the application execute?
2. Where is the data stored?
3. Can it function off the network (mobility)?
Using those criteria, the major compute models we track split out like this
I'll admit to a couple flaws in this picture. I change definitions for Blade PCs depending if they are used in 1-user-per-blade mode, or many-users-per-blade. In the 1:1 mode, it is just a repartitioned PC, so I classify them as "client-side". In 1:Many mode, the blades are more like servers. Web-based apps are also not pictured. I think of server-based web apps are similar to terminal services. Web apps that rely on client-side execution resources, such as Java, Flash, Ajax and Silverlight, behave like streamed applications.
Devices versus Compute Models
"Thin client" is a loaded term. To some, it's a software architecture. Others see it as a specific device - a terminal. I'm always careful to separate conversations about which compute model is appropriate for a situation from the client device. A desktop or laptop PC can act as a terminal or "thin client" when using server-side applications. My recommendation is first determine the compute models a user will need, and then figure out the right device.
The Ups and Downs of the Models
The list of pros and cons for each model is long, and many authoritative sources and pundit sites are available to debate them. I'll provide a quick list of what we see as the major benefits and biggest limitations of each.
Terminal Services
| Benefits | |
Limitations |
| Well-understood, mature software and management tools | |
No off-network mobility |
| Centralized application management and data security | |
Performance issues with any motion graphics or video |
| Greater clients-per-server ratio than other server-side models, making it lower cost to deploy | |
Compute load, graphics, number of users and distance can all impact user experience |
| Stateless clients for simple adds, moves and changes | |
May be unable to use or synchronize local devices or peripherals (PDAs, barcode readers, digital instruments, etc.) |
| Access to applications from any network connected PC | |
Critical failure points in server and network that can affect all users |
Virtual Hosted Desktop
| Benefits | |
Limitations |
| PC-like user experience with GUI and application customization | |
No off-network mobility |
| Centralized management of complete desktop image, including OS and applications | |
Performance issues with any motion graphics or video |
| Centralized data security | |
Compute load, graphics, number of users and distance can all impact user experience |
| Stateless clients for simple adds, moves and changes | |
Lower clients-per-server ratio than other server-side models, making it higher cost to deploy |
| Access to applications from any network connected PC | |
May be unable to use or synchronize local devices or peripherals (PDAs, barcode readers, digital instruments, etc.) |
| |
|
Management tools not as evolved as more mature compute models |
OS + Application Streaming
| Benefits | |
Limitations |
| Centralized management of complete desktop image, including OS and applications | |
No off-network mobility |
| Centralized data security | |
Network must have sufficient capacity to handle "boot storms" |
| Application responsiveness due to local execution, even for graphics, audio and video | |
Does not remove requirement to tune image to specific client hardware configs |
| Stateless clients for simple adds, moves, changes and software fixes | |
Sequencing process can be time and labor-intensive |
| Fewer servers required to service a population of clients versus server-side compute models, which lowers costs | |
Applications may not interact with each other as expected if virtualized. Additional engineering may be required to find optimal app combinations to bundle and virtualize together. |
Application Streaming (aka Application Virtualization)
| Benefits | |
Limitations |
| Centralized management of applications while providing an on-demand user experience | |
OS still installed locally, so IT must engineer processes to manage, patch and update distributed, local images |
| Data storage flexibility. Options for local or centralized storage, based on policy | |
High volume of simultaneous downloads may decrease application streaming speeds and network responsiveness. |
| Application responsiveness due to local execution, even for graphics, audio and video | |
If using local data storage, data may be at risk if not protected with passwords and/or disk encryption. |
| Off-network mobility with local caching options | |
Sequencing process can be time and labor-intensive |
| Fewer servers required to service a population of clients versus server-side compute models, lowering costs | |
Applications may not interact with each other as expected. Additional engineering may be required to find optimal app combinations to bundle and virtualize together. |
| By virtualizing and abstracting the application from the OS, many incompatibilities can be avoided, and greater application isolation can increase security | |
|
The Intel Point of View
After laying out these options and the pros & cons, customers inevitably want to know where Intel comes down on it all. Our view is this:
- In general, Intel believes the client-side models (application or OS streaming) provide the best balance of user experience, centralized management and data security. We believe they have the broadest potential to serve both users and IT managers well.
- Server-side models (terminal services, virtual hosted desktops) are appropriate and advisable in certain cases and applications. These are situations where mobility is not required, application workloads don't call for motion graphics or security requires tight lock-down.
- In reality, users will access applications via more than one compute model. A user may use locally installed office applications, but access databases through terminal services. A student may access the standard department image with a virtual hosted desktop, but use a local browser and media players.
- If using a client-side model or a mix of models, a PC is the best client device, particularly those with Intel^®^ vPro^^TM technology. For pure, server-side deployments, terminals are appropriate, driven by 64-bit, multi-core Intel^®^ Xeon^®^ Servers.
Let the Games Begin
We are living in interesting times. Evolution of computing models is happening in real-time, and there is no shortage of opinions and viewpoints. I hope this blog has been informative. Even if it hasn't, I hope it is a launching pad for a spirited (yet respectful) debate here. Let the games begin!
