Forgot your password?
typodupeerror
The Internet Hardware News

Amazon's Cloud May Provision 50,000 VMs a Day 122

Posted by kdawson
from the golden-lining dept.
Dan Jones writes "It has been estimated that Amazon Web Services is provisioning some 50,000 EC2 server instances per day, or more than 18 million per year. But that may not be entirely accurate. A single Amazon Machine Image (the virtual machine) may be launched multiple times as an EC2 instance, thereby indicating that the true number of individual Amazon servers may be lower, perhaps much lower, than 50,000 per day. So, even if it's out by a factor of 10 that's still 1.8 million VMs per year. Is that sustainable? By way of comparison, In February of this year, Amazon announced S3 contained 40 billion objects. By August, the number was 64 billion objects. This indicates a growth of 4 billion S3 objects per month, giving a daily growth total of about 133 million new S3 objects per day. How big can the cloud get before it starts to rain?"
This discussion has been archived. No new comments can be posted.

Amazon's Cloud May Provision 50,000 VMs a Day

Comments Filter:
  • If history tells us anything, it is that there _will_ be a failure.

  • tag: Dumbquestion (Score:4, Insightful)

    by drinkypoo (153816) <martin.espinoza@gmail.com> on Wednesday September 30, 2009 @08:27AM (#29591773) Homepage Journal

    How big can the cloud get before it starts to rain?"

    Clouds don't work like that, they let go their rain when they enter a pressure zone where they can no longer hold water.

    If Amazon is centrally dispatching, then they deserve to fail. If not, then there's no reason why getting larger would necessarily cause any particular problem.

    • Re: (Score:3, Interesting)

      by Enry (630)

      This. Maybe instead of atmospheric clouds, they're talking about the Oort Cloud [wikipedia.org].

      • Those comet thingamabobs don't rain now do they?
        • Re: (Score:3, Funny)

          by Enry (630)

          If EC2 has the same uptime as bits of that cloud destroying life on earth, I think it'll be around for a while.

          And if one does hit us, I guess it won't matter anyway.

  • Please stop... (Score:5, Insightful)

    by broken_chaos (1188549) on Wednesday September 30, 2009 @08:28AM (#29591783)

    Cloud is bad enough. Starting up bullshit analogies with clouds and rain just muddy whatever you're talking about far, far more than is necessary.

    • by Josh04 (1596071) on Wednesday September 30, 2009 @08:34AM (#29591855)
      I agree, the rain does muddy the waters somewhat. Not to mention the flood of comments deriding it as such.
    • Yes the summary didn't make any sense to me either. "How big can the cloud get before it starts to rain?" Huh? Is it saying how long will it be until the cloud starts making a profit for Amazon, or until the cloud collapses under its own weight? I still remember when AOL signed-up too many customers, and the result was a service that was slow and unresponsive.

      • Re: (Score:3, Funny)

        by smaddox (928261)

        I still remember when AOL signed-up too many customers, and the result was a service that was slow and unresponsive.

        Yeah, I remember their grand opening, too.

        • by Zerth (26112)

          And the year after that, and the year after that, and ...

          • Actually AOL was a perfectly good and usable service for its first ten years of online existence (1985-95). It wasn't until the pressure of tens of millions of newbies joining AOL that the servers were found to be inadequate for the task. I suspect we'll see the same will be true with Amazon and other cloud companies - good at first but then experiencing slowdown as their central servers become overwhelmed.

        • by mcrbids (148650)

          Yeah, I remember their grand opening, too.

          Somehow, I doubt it, since it was in 1983....

      • AOL did that several times as I recall, seemed they never learned to scale properly. I also remember that they sent out so many floppies back when they were still useful, that I never had to buy any...I often wondered how they could send that many floppies out in pretty packaging and stay in business...later they started using cdroms, so then they became a nuisance as apposed to something useful.
        • Re: (Score:3, Interesting)

          by Gilmoure (18428)

          Oh man, I was in art school in early 90's. All those AoL CD's were great for material for art projects and stuff.

        • by Ash-Fox (726320)

          later they started using cdroms, so then they became a nuisance as apposed to something useful.

          Why didn't you build a AOL CD throne [bedzine.com] like the rest of us?

    • by rotide (1015173)
      Cloud is indeed bad enough. But then talking about the mud that is caused by the rain just entirely washes away any sense it could have made.
    • by suso (153703) * on Wednesday September 30, 2009 @09:04AM (#29592153) Homepage Journal

      Oh, stop raining on everyone's parade.

    • Re:Please stop... (Score:5, Insightful)

      by moon3 (1530265) on Wednesday September 30, 2009 @09:42AM (#29592601)
      Managers love this kind of terminology, because from their point of view Internet just 'happens' somehow, they do not have a real clue how, but the cloud fits perfectly into this kind of thinking. That is why cloud hosting is so popular, they just order 4GB/100Mbit/s cloud and the hosting company creates one for them. They do not have to worry about setting up DNS, SQLs, multiple servers, domains, SMTPs and get schooled by some lowlife nerdy IT guys, they understand the dumbed down cloud interface well enough themselves, they just interact with the web interface and are happy it is all working for them.. somehow, somewhere, in the cloud.
      • by hodet (620484)
        This is so well put and describes where I work perfectly.
      • Re:Please stop... (Score:5, Informative)

        by slim (1652) <john AT hartnup DOT net> on Wednesday September 30, 2009 @11:02AM (#29593757) Homepage

        Managers love this kind of terminology, because from their point of view Internet just 'happens' somehow.

        And cloud computing makes them right. You pay some money, and the entity you're paying the money to, makes it happen.

        Just like when I buy a tin of soup from a supermarket, I don't need to understand anything about the supply chain that got it there.

      • by identity0 (77976)

        I don't know about you, but the networking class I took 5 years ago used the term "cloud". Both the teacher and the textbook used it to refer to networks like the internet, which you send data through but are outside of one's control and whose layout is not relevant to you.

        I think "rain" might be taking the analogy too far, though.

    • Wait 'til you see my value-added cloud service that adds cloud seeding and piss recycling.

  • who cares how many potential VMs the "cloud" can host. its methodone for most end users/devs real problems: inefficient code. the "just pitch machines at it until it runs fast!" mentality will catch up to us.
    • by RealityProphet (625675) on Wednesday September 30, 2009 @08:36AM (#29591877)

      who cares how many potential VMs the "cloud" can host. its methodone for most end users/devs real problems: inefficient code. the "just pitch machines at it until it runs fast!" mentality will catch up to us.

      That's not true. We use Amazon's cloud to host some of our servers. The reason we do it is for two main reasons. (1) We don't need to worry about equipment maintenance. Let me repeat that lest you think its not a big deal: We don't need to worry about equipment maintenance! (That is a big deal when you leave your basement but don't necessarily have a dedicated IT staff). (2) We are in a rapid growth phase. We cannot estimate well enough what are computing needs, our storage needs, are going to be 1- 2- 6- months down the road. We also don't have $50k to drop on equipment and storage that may be utilized 6 months from now, but we sure as hell know if we bought it now it wouldn't be used immediately. Amazon's cloud makes it trivial to keep up with our growing demand without paying up front for it. Sure we pay more to "rent" the stuff from Amazon, but its simply the big(O) argument: Amazon's pricing scales worse than the classic alternatives, but the constants out front are tiny.

      • by commodore64_love (1445365) on Wednesday September 30, 2009 @08:46AM (#29591975) Journal

        So to use a car analogy (cough)

        - It's the same reason why people lease cars instead of buying them. It's cheaper in the short term, and easier to come up with $300 for rent than $20,000 for purchase. Plus adding extra cars as new employees join the company is trivially easy.

        • by afidel (530433) on Wednesday September 30, 2009 @09:45AM (#29592655)
          But it's even better than a car lease, because you can end the lease on the VM with no penalty. If you have a really big batch job that needs to run once a month then you just spin up the VM's for the duration of the batch job paying for your usage and them deprovision them for the rest of the month.
        • by alen (225700)

          not really

          a lot of businesses don't have cash on hand to meet their needs even if they are profitable. Even Best Buy has to borrow short term to buy up enough inventory to meet demand. Suppliers want to be paid right away. Amazon's solution is ideal for cash poor companies

        • by maraist (68387) *
          To further your point. There is no such thing as free money.. Even if you have $50M in the bank, you don't necessarily have 'free money'.. There is something called the opportunity cost of holding money. And also a liquidity-demand curve. Namely, you can put that dollar into a CD, or commercial paper or a bond or whatever. And your CFO SHOULD be doing this with every penny the company has (with a tapering liquidity curve). So if you can borrow money cheaper than you can safely invest it, then you lite
          • And sometimes it can lead to stupid decisions, like when my company leased a Tektronix Logic Analyzer for $5000 a year for six years. They could have just bought the thing for $20,000 at the start of the project and saved money.

      • We don't need to worry about equipment maintenance.

        For the scenario you described, I think S3 would be a good choice. Likewise if a bigger company had a division or department with out-sized or highly variable data storage needs, might work in that situation as well. Judging by the number of objects, a lot of people are finding uses for that capacity.

        I know for a while Walmart was using some paint-by-numbers hosted application provider that was based on ASP. Don't know if they still do, but for those

  • by Viol8 (599362) on Wednesday September 30, 2009 @08:33AM (#29591827)

    I've never really understood the fuss around VMs. Sure , they're useful if you want to test run an OS install or run a different OS on top of another. But otherwise whats the point? Instead of having app + OS you end up with app + VM + OS so how exactly is that benefiting anyone other than the power company for the extra electricity used?

    • If you have a 4 core cpu you could run 4 VMs on one machine at one core per VM.
      • by reashlin (1370169) on Wednesday September 30, 2009 @08:52AM (#29592039)
        Its more than that.

        Most machines run at around 10% of all possible utilisation. Often web servers will run at less than this. In a datacenter you have two options a) run hundreds of very slow cheap machines each running one instance of your webserver. b) consolidate lots of machines onto one powerful box and running it at 70-80% utilisation.

        Option b) has the advantage that should a website get hit heavily (maybe because its been linked too on /.) then you still have the beefy hardware to cope with it. You will also find heating bills go down. You'll usually even get the costs of the hardware down as well.

        If your still not convinced then look at the work by most VM software manufacturers who are making it so the VM can move around on physical hardware. Now if your hardware fails - the VM and OS does not. It just moves off somewhere else and continues to operate with little/no drop in performance or uptime.
        • look at the work by most VM software manufacturers who are making it so the VM can move around on physical hardware. Now if your hardware fails - the VM and OS does not. It just moves off somewhere else and continues to operate with little/no drop in performance or uptime.

          This is the point where our sci-fi future is already visible as a thin shape on the horizon. A VM that lives independently of hardware points toward future abstraction layers where everything is machine-independent, managed transparently b

          • by Phurge (1112105)
            I've just finished Neuromancer.... " Cyberspace. A consensual hallucination experienced daily by billions of legitimate operators, in every nation, by children being taught mathematical concepts. ... A graphic representation of data abstracted from banks of every computer in the human system. Unthinkable complexity. Lines of light ranged in the nonspace of the mind, clusters and constellations of data."
    • Re: (Score:1, Informative)

      by Anonymous Coward

      Because your apps shouldn't be using 100% of the power of the machine they're on. With virtualization I can meet all of my hosting needs on less net hardware, even after you take out the inefficiency of the virtualization.

      • Re: (Score:1, Troll)

        by Viol8 (599362)

        Sorry , that makes no sense. By definition you could do it on the same hardware without a VM unless your VM somehow magics processing power out of the ether.

        • by hodet (620484) on Wednesday September 30, 2009 @09:58AM (#29592825)
          It makes perfect sense. His clients want a dedicated host for their server. 10 clients, 10 virtual servers on one powerful box instead of 10 servers running at minimum capacity. More profit for parent. Data Centers are using virtualization big time because it saves money. Very easy to move the guest OS around if needed, even geographically.
        • by bertok (226922) on Wednesday September 30, 2009 @10:29AM (#29593223)

          Sorry , that makes no sense. By definition you could do it on the same hardware without a VM unless your VM somehow magics processing power out of the ether.

          Except that unless you have a magic crystal ball, you'll never be able to predict application load ahead of time. Hence, some servers will be underutilized, and some will be sitting at 100% half the time. The only alternative is to install every application onto every server you have, and load balance everything - but that requires that every app is compatible with every other app, and that every app can operate as a cluster. In practice, that's impossible for typical businesses.

          What the latest virtualization platforms do is load balance, on the fly. A large VMware cluster will analyze the load pattern and redistribute virtual machines around the cluster to balance things out, so that each host is evenly loaded. I've seen clusters set to an average of 70% CPU load, and it was just fine. If one host starts heading towards 100%, a few VMs are shuffled around until the load is evened out again. Users can't really tell the difference between, say, 20% and 70% load. It's only at 90% or higher that you get contention and increases in response latency. It takes about 5 seconds to move a VM, but the actual outage is only a few milliseconds, if that, so users never notice.

          One thing I noticed with VM deployments is that most apps get faster on less hardware. This is counterintuitive, but I've seen it before in well designed Terminal Server / Citrix deployments. The basic concept is that you can afford much better hardware if you need less of it. You can buy beefier servers, 10Gb ethernet, SAN storage, etc... When 1 app needs lots of power, it gets it, and then it gives up its share when it doesn't to other apps that do.

          So yeah, in a sense, virtualization does magic processing power of the ether, because it actually lets you use the processing power you paid Intel or AMD thousands of dollars for.

    • by SappoMan (51574) on Wednesday September 30, 2009 @08:47AM (#29591987)
      Ok, you don't work in IT right? At least not on the admin side.
      VM are mainly about server consolidation. That means that given the fact that servers are usually under utilized you can put quite a number of VM per core. Usually for server workloads the number is around 2: 2VM * 4 cores * 2 cpu (typical blade) yields 16 VM. You see, in the end the power company gets paid only for a physical server every 16 OS instances. Not bad.
      Server consolidation is not the only reason you use virtualization. Other issues you can solve are: high availability and fault tolerance, quick deployment of new servers, hardware abstraction and many others
      • by felipekk (1007591)

        Hardware abstraction being one of the more compelling features IMO.

        How long you think it would take you to move all the services to that new server hardware you just received, because the current server hardware warranty just expired?

        Well, with VM's you can do that _A LOT_ faster.

        • by SappoMan (51574)
          Yes, between 10 seconds and a minute.. ;) left click, migrate and BOOM
        • by Vancorps (746090)
          Takes about 10 seconds and happens without dropping so much as a packet. Of course that will depend on your back-end storage. I'm fortunate enough to run with NetApp though so it's all gravy. On a side-note, Citrix Essentials for XenServer is pretty cool with their SAN integration technology. Thin-provisioning on the fly is the way to be! When I make a storage repository it is 20megs until I create a VM and even then it will only take up space that is actually taken up. Combined with volume level deduplicat
          • by slim (1652)

            Just wish people would stop calling it a cloud.

            What you have might not be a cloud, but what Amazon offers probably is.

            You provision an EC2 instance, and it's there. You don't know where it is, and you don't care. There's just a cloud of VMs, yours is in there somewhere, and you can connect to it to use it.

            Dynamically, Amazon might choose to move your VM to another physical host for all kinds of reasons -- an unplanned physical failure, a planned power-down, some algorithm that decides they could meet your needs more cheaply at a different location, etc.

            • by Vancorps (746090)

              That is precisely what you get with Citrix XenServer combined with Citrix Essentials, there are many other products for Xen that give you everything that Amazon is offering including virtual switches and load balancers.

              When I want new instances I just spawn new instances and it will take care of provisioning storage for me. I'll grant it's not as slick but there is no reason to call it the cloud when all it does is virtualize all your infrastructure. You're not using any new software on the back-end to cre

      • Re: (Score:1, Troll)

        by Viol8 (599362)

        I thinl you're missing my point - why have multiple OSes if they're all the same type of OS and the apps could all happily run on the same OS instance? As for deployment - have you never heard of a tarball? OS dies - take app tarball to new server , untar. Hows that different to copying a VM machine file over?

        • by bertok (226922) on Wednesday September 30, 2009 @10:09AM (#29592971)

          I thinl you're missing my point - why have multiple OSes if they're all the same type of OS and the apps could all happily run on the same OS instance? As for deployment - have you never heard of a tarball? OS dies - take app tarball to new server , untar. Hows that different to copying a VM machine file over?

          In the real world, people run apps like Exchange or Oracle, which take hours to install to a vanilla state, and that's not counting the potentially terabytes of data associated with them.

          Even the most primitive "tar ball" Linux app will have dependencies on the OS, and those can and will eventually break, unless you freeze your OS version forever. If you have enough apps and servers, that will become a nightmare to manage. Do I upgrade or not upgrade? Will this patch or that patch break one of the apps? This is how people end up running Linux 2.2, or 32-bit Windows on 64-bit platforms, because migrating 1 app is hard enough, but migrating a server with 20 apps on it is a recipe for disaster.

          Virtualization lets you quite literally drag & drop a running host OS from server to server. During maintenance time, that's like magic. No more 3am hardware replacement jobs for me! You can clone a machine while it's running, isolate the clone onto a virtual network, and test an upgrade without interrupting users. Sure, you can do that with most backup & restore tools, but VM platforms do it quicker, and with fewer admin steps. You don't even need spare hardware.

          I once replaced every single hardware component of a running VM farm, servers, cables, switches, even the SAN, while it was running. During the day. Zero outage, no packets lost, no TCP/IP connections closed or user sessions disconnected. We even had terminal server (Citrix) and console (SSH) users on. Not one user even noticed what was going on. I'd love to see you try that with 'tar'.

          • Handy for upgrades but also great for redundancy.

            If you have a power outage in a data center you can have all of the servers instantly pop up in the backup data center without even dropping any sessions.

          • FWIW, Oracle takes about 30 minutes to install to vanilla state & create a basic database.

            The big problem with VMs is image proliferation. Package-level management, like RPMs, etc. still is essential to understand versioning dependencies; virtualization does not fix that problem.

        • by dissy (172727)

          why have multiple OSes if they're all the same type of OS and the apps could all happily run on the same OS instance?

          Well if that situation was ever true, you might have a point.
          In the real world, it rarely works out that way.

          Especially so for load balancing.

          Your solution is similar to a server with 2-3 power supplies, all plugged into the same powerstrip, yet claiming redundancy. It just doesn't work like that.

        • by dissy (172727)

          As for deployment - have you never heard of a tarball? OS dies - take app tarball to new server , untar. Hows that different to copying a VM machine file over?

          Sorry for the second reply, but...

          I would imagine a tarball of an OS would take more than 2 seconds to move from one machine to another.
          Every second over 2 seconds is wasted time.

          Yes, you can move a running VM from one physical server to another, in seconds, while the OS in the VM never notices anything but a 100ms network burp (if that)

          When you can take a server down, tar its backup to a new machine, boot it, and get it running, without any of the apps being offline at all nor the guest OS restarting, then

        • by Zerth (26112)

          I thinl you're missing my point - why have multiple OSes if they're all the same type of OS and the apps could all happily run on the same OS instance?

          HIPAA, Sarbanes-Oxley, and a few credit card companies would like to have a word with you.

      • Re: (Score:3, Insightful)

        by timeOday (582209)
        The point is that multi-tasking operating systems already support server consolidation by protecting processes from each other so you can run multiple processes on a host safely. And they do it in a FAR more efficiently than VMs, which have an entire OS instance for every process, and memory partitioned statically between them.

        However, the OS doesn't quite finish the job. The need for VMs arises from design shortcomings at the OS level and above. Here are a few:

        1. You can't install an app and all its dep
    • Re: (Score:3, Informative)

      by alen (225700)

      we have used VMWare for a few years. Our devs would write a java app and it would require it's own server but it would use maybe 20% if not less of the resources. Now we just provision a VM. less server clutter in the datacenter and smaller electricity bills. Also great for DR. we ship the entire VM to a DR site so all we have to do is bring it up, change the IP and we're ready to go. otherwise we would spend days trying to configure all the apps, find the source, etc.

      i have my own server i used to test a S

      • by radtea (464814)

        Our devs would write a java app and it would require it's own server but it would use maybe 20% if not less of the resources.

        This is the part I don't get, that is left out of the answers above (the migration issue makes sense independently, though!)

        My question is simple: how on Earth do you write an app that "would require its own server" but only use 10 or 20 percent of the machine's resources? I Just Don't Get It when you say an app would "require its own server" but not max out the server's resources.

        W

        • Re: (Score:3, Informative)

          by jcnnghm (538570)

          Security/Separation of Duties.

        • Re: (Score:3, Interesting)

          by alen (225700)

          having one app conflict with another app. 10 years ago we had a few apps. today there are too many to count and constant point releases where minor functionality is added by user request or small bugs fixed.

          and it's not just java apps. weblogic instances, other apps we might buy or code internally. then there is QA since they need everything production has. Moving QA to VMWare was one of the first things we did when we bought it. the QA and Dev SQL servers are still physical, but a lot of their apps are now

    • VM's are great for many things. First off, know that most hardware is severely under-utilized. Then factor in the ease of replication, testing, security(via sandboxing and other methods), ability to scale horizontally quickly. There are downsides too of course which is why we prefer to run our own XEN setup, then use http://www.eucalyptus.com/ [eucalyptus.com] light up more VM's in case of load need or disaster.

      VM are a huge cost saver, and the fastest development environment.

    • by Synn (6288)

      Change "OS" to "hardware", so it's:

      app/OS + hardware vs app/OS + VM + hardware. The fuss is you get to disassociate your app and OS from a specific piece of hardware. If the hardware fails all you have to do is move the VM "image" to new hardware.

      Or, if the needs of the up go up or down you can move it to less powerful(cheaper) or more powerful(expensive) hardware as needed without much effort.

      • by Viol8 (599362)

        No , that would be app/OS + VM + OS + hardware. The host OS doesn't just conveniently vanish, its still using up CPU cycles.

        "If the hardware fails all you have to do is move the VM "image" to new hardware."

        How many apps are hardware dependent these days? All OS's have hardware abstractions , eg unix /dev directory. So whats different between doing that and just moving the app over to another server?

        • by slim (1652)

          No , that would be app/OS + VM + OS + hardware. The host OS doesn't just conveniently vanish, its still using up CPU cycles.

          While you can run VMs on a desktop OS (or indeed on a server), the real fun comes when you run it on a platform that's been custom designed for the purpose. For example, VMWare ESX. This is effectively a kernel with VMWare built into it. Sure it's an OS, but it's a very, very light OS.

          "If the hardware fails all you have to do is move the VM "image" to new hardware."

          How many apps are hardware dependent these days? All OS's have hardware abstractions , eg unix /dev directory. So whats different between doing that and just moving the app over to another server?

          The difference is that you can do it without stopping the application.

          Imagine you're paying me to host a Linux system for you. I could be running that as a VM on box a. While you're logged in and actively using the system, I c

        • by Synn (6288)

          How many apps are hardware dependent these days? All OS's have hardware abstractions , eg unix /dev directory. So whats different between doing that and just moving the app over to another server?

          Because you have to configure and setup that server for the app.

          I run 70 instances(servers) on Amazon EC2 with maybe 10 different applications/products. All of these instances are built from "images" called AMI's. When I update an instance, like say install a new gem(Ruby on Rails library), cron job, supporting perl script or whatever, I always re-image the server so the AMIs are up to date.

          When an instance dies I can build a new one from that image in about 5 minutes, all from my home or office. If I had t

        • by Ash-Fox (726320)

          How many apps are hardware dependent these days?

          I am aware of issues with software like Microsoft Exchange which will corrupt mail databases if it gets a I/O wait error (drivers aren't supposed to do this, but common in a lot of drivers) rather than the driver just pausing all I/O for the hardware involved to keep up. Running Exchange in a VM gets around this annoyance.

          Then there is hardware that doesn't support certain OSes correctly, (ie: win2k) which maybe the only OS a certain product will run on (due t

    • Basicly you're right.

      But there are some neat tricks you can do with VMs like taking an instant snapshot and use that for debugging.
      Migrating VMs to another (hardware) server is a non-issue. (just Copy over the image)
      If you're working with a cluster anyway, creating another node is also mainly a matter of copying the image.

    • by slim (1652)

      There are plenty of reasons why you might choose to host two services on two different machines, even if one machine would have enought power. Things like being able to take one down without affecting the other.

      VMs let you keep some of that model, while consolidating down to less hardware.

      Plus it makes deployment easy: get your system how you want it, then save it as an image. Now you can clone it as much as you like. Now that there are OSS VM hosts, the commercial virtualisation companies are concentrating

    • by teshuvah (831969) on Wednesday September 30, 2009 @09:04AM (#29592149)

      I've never really understood the fuss around VMs. Sure , they're useful if you want to test run an OS install or run a different OS on top of another. But otherwise whats the point? Instead of having app + OS you end up with app + VM + OS so how exactly is that benefiting anyone other than the power company for the extra electricity used?

      Because for the most part, most servers don't run anywhere near full capacity (and if they do, then they are probably not good candidates for virtualization, except possibly for high availability purposes which I will go over in the second paragraph). I forget the study but I read once that on average a typical server sits at 5-15% utilization. So the idea behind products like VMware ESX is that if you need 5 unique servers, instead of buying 5 servers at $5,000 a piece, you buy 1 server for $5,000 + 1 $5,000 VMware license, and run the 5 virtual servers on that. So you spend $10,000 instead of $25,000, and your footprint is 1/5th of what it was before, meaning less racks, less cooling, less power, etc. And the numbers I gave are very conservative. A lot of people do 10-20 VMs per server easily.

      So cost, power, and cooling issues aside, there are other issues. In a typical server environment, if a physical server suffers from a catastrophic hardware failure, that server is down until someone can walk over and swap the hardware. With VMware, if a VM is running on a server and that server fails, the VM is cold booted on another ESX server automatically, and is typically up in 30-60 seconds. With the newest release of ESX server, called vSphere, they take it a step further. You can optionally choose to have A VM mirror itself on to another physical ESX server. So in the event of a hardware failure, the VM keeps running on the mirrored host. And then, it becomes the primary VM and sets itself up to mirror automatically on another ESX server. So you have ZERO downtime and the app re-mirrors itself. These are just some of the many useful features in VMware.

      And no, I do not work for VMware. I am a contractor for the Air Force and over the past 2 years I have converted almost 200 physical servers to VMs. We are a relatively small program, but our projections show that we will save $2,000,000 over 10 years just on the cost of servers (and yes, i have added in the cost of VMware licenses and support into that equation), and that doesn't even account for power and cooling savings. We've gone from almost 200 physical servers distributed over 7 full racks racks down to 28 servers in 2 racks (2 racks only because they are two separate facilities. Each rack only contains a single HP c-class chassis)

      I think the real question is, how can you NOT understand the fuss around VMs?

    • There are two things that appeal to me about VMs.

      The first is the east of backup/recovery/spawning new VMs. Want to play with altering ProgramA, no problem, let me just copy ProgramA's VM and start it up.

      The other is less hardware. Perahps ProgramA and ProgramB don't want to run on the same server... they will generally run in seperate VMs fine. Perhaps ProgramA requires a different version of SQL, or some other dependancy; no problem in VMs. Things that were before going on underutilized servers can now be

    • I used them for two reasons.. Application Isolation, and Disaster recovery..

      Many of the apps we used (hello, Oracle Colaboration suite, looking at you) require really messing with system files to make work decent. This makes other programs very unhappy, so apps like these really need to run on their own box. Since it wasn't disk or CPU intensive, it was easy enough to just stick in a VM, so I could do other things with the server too. Secondly, its kinda nice when you need to restart a machine to fix a p
      • by radtea (464814)

        Many of the apps we used (hello, Oracle Colaboration suite, looking at you) require really messing with system files to make work decent. This makes other programs very unhappy, so apps like these really need to run on their own box.

        This makes sense now: incompetent server developers are the driver behind this aspect of VMs. Other aspects are independent of this, but a lot of people responding here have simply taken for granted that most servers are written by monkeys, and are therefore unable to play nic

    • by spec8472 (241410)

      We're a dev shop. Having a VM farm has saved our IT guys so much time and money.
      We used to buy dedicated boxes - some projects would get their own, most would share on some conglomerate box. Weird shit happened, stuff would never get uninstalled after a project was over, and people would be tripping over each other all the time (One person needs to reboot a box, the other was trying to debug some arcane issue). In short: a nightmare.

      Now, we buy a AUD$6k box from $brand, hook it up to our SAN, and run anywh

    • by sorak (246725)

      Someone mentioned server consolidation. Another purpose is that you can have each application run in it's dedicated environment. This reduces compatibility issues, provides flexibility, and has been used in one business for better load balancing in the event of equipment failure.

      Let's say you have an email server and a web server on a single machine (small organization). If you need a new server, and had set this up without virtualization, then it may be difficult and time consuming to rebuild the new machi

    • Net admin can tell you lots of benefits, but IMHO everything boils down to:

      VM is able to move from one machine to another but OS alone is not. All other benefits are either derived from this or possible to do without VM. Lets not forget that VM brings performance penalty. If Phoronix benchmark is valid, penalty is significant. So, ability to move running Linux OS (for example) around across machines would bring even more consolidation (more virtual servers running on single Linux machine as separate use
    • by dissy (172727)

      Instead of having app + OS you end up with app + VM + OS so how exactly is that benefiting anyone other than the power company for the extra electricity used?

      I think you have that backwards.
      How can using 8-10 TIMES less electricity be considered a benefit for the power company, whom makes more money the more electricity you use?

      If you need 10 machines, and run 10 servers, all 10 are using power.

      If you replace all 10 with just one server that uses 1/10th the power of the above setup, and run 10 VMs on it to replace those 10 real computers, you've just removed a decimal place from your power usage!

      • by Viol8 (599362)

        Or you just run 1 server with 10 apps running on it. Which I'm pretty sure is how things used to be done before Vms became flavour of the month amongst lazy sysadmins.

        • by dissy (172727)

          Or you just run 1 server with 10 apps running on it. Which I'm pretty sure is how things used to be done before Vms became flavour of the month amongst lazy sysadmins.

          How do you pull that one off when one or two of those apps won't run on anything newer than NT 4.0, a few only run XP or newer, and one wierd app requires windows 2003 server or better (but still server OS) and the other half of the applications are UNIX?

          Do you just install to multiple C:\windows_XXX\ dirs and reboot frequently between them?

          • by Viol8 (599362)

            So you're going to have 10 apps that all require a different server OS? I don't think so. At least not if the CTO has a clue.

            • by dissy (172727)

              So you're going to have 10 apps that all require a different server OS? I don't think so. At least not if the CTO has a clue.

              Well yea, OK. You are technically correct (the best kind of correct!)

              On one hand, that situation should NOT be as common place as it actually is, if things were done right.

              But on the other hand, most CTOs do not have a clue. Even the ones who do, unless they are hired in from the start, have good odds of having legacy crap to support.

              The 'newest latest and best' is not a phrase one usually associates with most business IT operations. Legacy, cruft, constant migrations, and CFO's calling the shots on tech

    • Wow, you must not run any servers.

      Just one very simple example of why VMs rule: hardware upgrades.

      We have an OS running outside of a VM, with tons of changes made to the setup, programs installed, etc. It's about 4 years old now. If I want to move all that stuff to hardware that's new, I am looking at starting up from scratch, configuring and installing all apps, and restoring data from a tape backup or transfer over the network to the new box.

      Compare this to running on a VM:

      (1) Get new machine
      (2) Put basic

  • by nweaver (113078) on Wednesday September 30, 2009 @08:41AM (#29591927) Homepage

    Lets give a 12 hour lifespan, and say 25K VMs at the same time.

    At 5 VMs/physical host (I suspect it is MUCH denser actually), thats only 5K servers. At 50 servers/rack, its 100 racks.

    Or, in translation, not THAT much.

    • How can you say 100 racks per day is "Not that much"?! Imagine provisioning 100 racks every single day! -- what about cabling and power and all that stuff? What about network load -- never mind the basic data transfer during normal operation, each server loads its disk image from the SAN when it starts. These seem pretty amazing numbers to me.
      • What makes you think that he did? If you are provisioning 100 racks a day it means that your total number of racks is increasing by 100/day. But if you read the subtlety in the GP's post

        Lets give a 12 hour lifespan, and say 25K VMs at the same time.

        then we find a subtle clue that these VMs don't live for ever. Admittedly we then have to understand his words to infer that they would only need 100 racks in total, rather than in addition every day. Go on, try it. Reading is fun. It's surprisingly informative w

  • I call shenanigans (Score:3, Interesting)

    by Anonymous Coward on Wednesday September 30, 2009 @09:02AM (#29592131)

    My company tried to provision 10,000 amazon instances to perform scalability testing of our software that runs on many computers. The math was simple - 10,000 servers * $0.15 / hour = $1,500 / hour for testing. We liked the multiple OSes & versions (Linux - Redhat, SLES, Windows - 2000, 2003, 2008?) and software stacks (mysql, apache, websphere, sql server, iis, etc...) that we all available out of the box.

    However, if you need more than 20 servers, you have to fill out a form. A sales rep and tech guy called to discuss our needs. It turns out that they could only handle around 1000 instance request across all data centers unless we "reserve" the machines at $300 / each, which blew the math - 10,000 servers * $300 = $3,000,000 to start.

    Looking at the article, it is likely that people are re-requesting the same machine be started & stopped multiple times per day - 50,000 is probably off by an order of 10.

    • by HappyDrgn (142428)

      "Looking at the article, it is likely that people are re-requesting the same machine be started & stopped multiple times per day"

      Guilty! I use EC2 as a cheap way to learn / test things in my spare time that I don't get to do at work. As a result I'm starting and stopping machines right now. In the next couple hours I'll probably launch a total of 10 instances for a couple hours.

  • Stock Exchange (Score:3, Interesting)

    by MyDixieWrecked (548719) on Wednesday September 30, 2009 @09:30AM (#29592453) Homepage Journal

    I went to an Amazon's AWS talk in NYC a couple months ago where they brought some start-ups in to talk about their projects, the cloud and how the cloud helped them build their applications faster and better. During the opening talk, the speaker showed some use-cases, one including the New York Stock Exchange and how, at the closing bell, they provision over 3000 EC2 instances to crunch numbers overnight to be ready for the next morning.

    A guy from a startup that I was talking to before we were seated was talking about how his company keeps between 5 and 10 instances up all the time for their application (dynamically bringing them up and down to scale with demand) and how they frequently had 4 and 5 sets of these servers running on the side for testing (20-40 instances at a time). He was talking about the metrics they were using to keep track of their use and how it was flawed due to the fact that they had hundreds of instances a day going up and down all the time.

    Just because 50,000 instances are started per day doesn't mean that those 50,000 instances are running for any period of time. I frequently bring up an instance, tweak some things, create an image, then bring it down... or bring up an instance to test something for 20 minutes, then bring it down. EC2 has really benefitted my QA/Testing/Experimentation in that I really have an unlimited pool of resources to play with. It's a much more robust system than I have at home with VMWare... vmware was a gamechanger for me since before that, I had 2 physical servers at home and stacks of 40GB and 60GB HDs with multliple versions of OSs on them.

    Of course AWS isn't for everyone. EC2 can be expensive for what they offer and the biggest advantage to AWS's services are that they are on-demand and work really well with applications that need to scale up AND down in real-time. If you've got an application that doesn't require to-the-minute scaling responses, it's less expensive to get a physical dedicated server with Xen on it and create your own virtual infrastructure... although if you don't have the skills or time to learn the tools, then AWS offers a much better learning curve.

  • Define "Objects" (Score:3, Informative)

    by Dersaidin (954402) on Wednesday September 30, 2009 @10:22AM (#29593119)
    Objects?

    "Objects" doesn't mean VMs, objects can be files, processes, etc.

    • by radio4fan (304271)

      In S3 terms, an object is effectively a file.

      It's an atomic blob of data up to 5GB in size, with up to 2KB of metadata.

  • Amazon EC2 is extremely attractive to spammers. They set up disposable spam servers (often using stolen credit card numbers). Amazon's policy is to shut down the account when this type of activity is detected, but they don't take any steps to keep the spammer from returning and setting up a new account. Based on this practice, the "50,000 VM's per day" seems quite reasonable when you consider the volume of activity that most spammers tend to produce.

panic: kernel trap (ignored)

Working...