2 posts tagged “virtualization”
Today I came across a question about virtualizing web servers and databases. It caught my eye because I'd asked a similar question back in November.
The question got me thinking about scalability. Whatever current woes running a database or any other application in a virtualized server may have now, they will eventually be solved or at least sufficiently mitigated to make it worth the penalty.
At this point we'll start seeing server applications that not only support virtualization, but require it. A virtual server can provide a standard environment that an operating system on traditional hardware doesn't quite match, and allows you to tune the resources available to a server without shutting down the machine. Imagine virtualization software that allows a live transfer of a system from one physical machine to another with no shutdown. This is already happening. What's interesting is where we go from here. I think the next logical step is clustering/load balancing and cloning.
These are features that are available now, but smaller businesses largely ignore. But I think that once all your major applications are on virtual servers, the clustering problem becomes much simpler. The load balancer will be a function of the virtualization software itself. Of course the individual application will have to support it as well, but this will become a core feature of any serious server application, even in the entry level editions. Setting up a cluster will be as simple as clicking a check box and typing in the cluster name. Maybe you won't even need to click the checkbox: you'll just give two or more servers the same dns name and the virtualization software will know what it means.
At this point, a small business with one server can add additional horsepower simply by buying a server (any server), plugging into the network, and telling their current virtual system to clone itself to the new machine. Large businesses can do the same thing.
For a practical example, let's look at the new StackOverflow servers. One database and two smaller web servers. Under what I'm proposing each of those physical servers would carry a database instance and a web instance, each in a three-server cluster. If one goes down, the other two keep chugging happily away. Of course, if the current database server were to go down the site would be hurting.
But to be fair the planning would be different under my proposal, and Jeff probably would have purchased four machines more like the web servers, rather than one database and two web servers. This way he ends up with about the same power for about the same money. Perhaps even more power because he can add more memory cheaply. One server failure isn't a big deal now.
Even more significantly, as Jeff's traffic increases it becomes very easy to scale up the site infrastructure to match. All he needs to do to add a server is drop it in the rack and tell the virtualization software to create a new instance and hook it up. As it is, he's looking at having to do a major overhaul of his infrastructure in about two years.
Virtualization has been heralded for a few years now as the 'next big thing' in IT. In fact, where I work they've nearly finished virtualizing the entire server room (~20 servers). Here's a look at what virtualization is, and how it can help a even small businesses.
To understand what virtualization is, let's start by looking at a server. You can thing of a server just like a regular desktop computer. It has resources like memory, a processor, and a hard drive, with an operating system that runs on top of all that. Now let's twist things up a bit. You take the server hardware and install a special operating system. This operating system provides a new base onto which you can install your regular operating system. Not only that, you could install several operating systems side by side-- all on one set of hardware. You gain an advantage this way because servers tend to be either way underutilized for simple tasks or way overburdened for complex or frequent tasks. You can now easily combine several underutilized servers in one place, freeing up hardware to share the load for the overworked servers.
Additionally, virtualization enables a number of new things that were not possible before, but I won't go into that here. I think the benefits for larger enterprises are pretty well documented. But what about smaller businesses? Here are two scenarios that show how virtualization can improve reliability and help promote growth for even the smallest businesses.
First, let's imagine you run a smallish business. You have between 15 and 50 computer users-- just big enough to need a full-time IT guy, or at least keep him busy enough he's cheaper than paying a consulting firm all the time. You probably have one server that acts as your domain controller, file server, and print server. You have another server running your web site, e-mail, intranet, and perhaps a custom business application. You have a third server running your database. This is a very common set up. Each of these servers is a single point of failure that can stop your business in it's tracks. Unfortunately, you're not big enough to really justify redundancy for all of those machines, but you do have a good UPS and overall they're pretty reliable. In fact, you run for months at a time with no trouble at all. However, when something happens, it happens big.
Enter virtualization. Now you set it up so that each server is a virtual host. Each host runs an instance of what used to run on each server, and they're configured to do load balancing and fail-over with the other two. Part of setting up the hosts is creating a virtual SAN so that the disks are shared in an efficient way. Now if a server goes down it's no big deal. More than that, as your business grows it scales up by just adding new boxes to the cluster- the hard work is already done. About the only thing you'll want to do later is add a real SAN to replace the virtual one.
This also adds flexibility to your setup with regards to deploying new systems. You can add servers for new features or test new OS or software versions in an isolated environment without having to purchase any additional hardware. You can take a system down for maintenance or upgrade during business hours without any loss of service.
Unfortunately, at this point getting the clustering to work is much easier said than done, but that is improving. And while still classified as a small business, the business in this example was still big enough to require a full-time IT staff member. So let's try another example.
Let's say I have a small software business. Since I hope to do that one day, this isn't far off. I have fewer than four users, but because I'm selling software I'm at the very least providing large downloads. I probably also need source control somewhere, and possibly provide a computer intensive service over the web. I also have a need for a testing environment that is guaranteed not to interfere with my production site. In short, I have processing needs that far outstrip what you'd normally need for this size user base.
In the past, to solve these problems I probably kept one server on site that's really more of a glorified workstation. This computer functions as source repository, pint server, and test environment. My web site, e-mail, and important backups are kept on a machine leased in a datacenter somewhere. This machine will automatically take care of backups, have a high quality internet connection, and have a qualified system administrator looking after it. If something goes wrong it the support contract ensures a replacement is up and running fairly quickly. If my needs grow beyond one production server, I can lease another machine
This setup is pretty good, but we can do a little better. I'll start by virtualizing the on-site server. Now I can isolate my test environment in it's own virtual machine. This will allow me to prevent my test system from bringing down my source control and print server. Even if I just do something that tries to claim all the resources I can protect myself by setting a quota. It also allows me to easily re-configure my test environment without interfering with the other system. In that past some people would have even required a second server to accomplish these same goals. Now I know I only need one.
Now let's look at the hosted server. There are two problems with this situation. First, you have to buy computing power a whole machine at a time. Secondly, occasional brief outages are still possible. These days, you don't have to lease a whole server. You lease a virtual server. You'll have to share the physical box with a few other people, but it's okay because you don't know who they are, they don't know who you are, and no one can see anyone's system but their own. You are protected from a peer grabbing all of the resources on the machine because your virtual server is actually running in a cluster on several machines. This also minimizes the danger of outages due to upgrades or individual machine failure. You save money because you pay a base rate for the virtual server, and then only pay for the computing resources you actually use. This could seem to get expensive depending on rates, but if you're using more computing power it's likely because you have more customers or sales leads, and this is a good thing. In fact, virtualized servers start for much less than leased physical servers.
In short, virtualization can save you money, increase reliability, and allow you to do more, no matter what size your business is.
To understand what virtualization is, let's start by looking at a server. You can thing of a server just like a regular desktop computer. It has resources like memory, a processor, and a hard drive, with an operating system that runs on top of all that. Now let's twist things up a bit. You take the server hardware and install a special operating system. This operating system provides a new base onto which you can install your regular operating system. Not only that, you could install several operating systems side by side-- all on one set of hardware. You gain an advantage this way because servers tend to be either way underutilized for simple tasks or way overburdened for complex or frequent tasks. You can now easily combine several underutilized servers in one place, freeing up hardware to share the load for the overworked servers.
Additionally, virtualization enables a number of new things that were not possible before, but I won't go into that here. I think the benefits for larger enterprises are pretty well documented. But what about smaller businesses? Here are two scenarios that show how virtualization can improve reliability and help promote growth for even the smallest businesses.
First, let's imagine you run a smallish business. You have between 15 and 50 computer users-- just big enough to need a full-time IT guy, or at least keep him busy enough he's cheaper than paying a consulting firm all the time. You probably have one server that acts as your domain controller, file server, and print server. You have another server running your web site, e-mail, intranet, and perhaps a custom business application. You have a third server running your database. This is a very common set up. Each of these servers is a single point of failure that can stop your business in it's tracks. Unfortunately, you're not big enough to really justify redundancy for all of those machines, but you do have a good UPS and overall they're pretty reliable. In fact, you run for months at a time with no trouble at all. However, when something happens, it happens big.
Enter virtualization. Now you set it up so that each server is a virtual host. Each host runs an instance of what used to run on each server, and they're configured to do load balancing and fail-over with the other two. Part of setting up the hosts is creating a virtual SAN so that the disks are shared in an efficient way. Now if a server goes down it's no big deal. More than that, as your business grows it scales up by just adding new boxes to the cluster- the hard work is already done. About the only thing you'll want to do later is add a real SAN to replace the virtual one.
This also adds flexibility to your setup with regards to deploying new systems. You can add servers for new features or test new OS or software versions in an isolated environment without having to purchase any additional hardware. You can take a system down for maintenance or upgrade during business hours without any loss of service.
Unfortunately, at this point getting the clustering to work is much easier said than done, but that is improving. And while still classified as a small business, the business in this example was still big enough to require a full-time IT staff member. So let's try another example.
Let's say I have a small software business. Since I hope to do that one day, this isn't far off. I have fewer than four users, but because I'm selling software I'm at the very least providing large downloads. I probably also need source control somewhere, and possibly provide a computer intensive service over the web. I also have a need for a testing environment that is guaranteed not to interfere with my production site. In short, I have processing needs that far outstrip what you'd normally need for this size user base.
In the past, to solve these problems I probably kept one server on site that's really more of a glorified workstation. This computer functions as source repository, pint server, and test environment. My web site, e-mail, and important backups are kept on a machine leased in a datacenter somewhere. This machine will automatically take care of backups, have a high quality internet connection, and have a qualified system administrator looking after it. If something goes wrong it the support contract ensures a replacement is up and running fairly quickly. If my needs grow beyond one production server, I can lease another machine
This setup is pretty good, but we can do a little better. I'll start by virtualizing the on-site server. Now I can isolate my test environment in it's own virtual machine. This will allow me to prevent my test system from bringing down my source control and print server. Even if I just do something that tries to claim all the resources I can protect myself by setting a quota. It also allows me to easily re-configure my test environment without interfering with the other system. In that past some people would have even required a second server to accomplish these same goals. Now I know I only need one.
Now let's look at the hosted server. There are two problems with this situation. First, you have to buy computing power a whole machine at a time. Secondly, occasional brief outages are still possible. These days, you don't have to lease a whole server. You lease a virtual server. You'll have to share the physical box with a few other people, but it's okay because you don't know who they are, they don't know who you are, and no one can see anyone's system but their own. You are protected from a peer grabbing all of the resources on the machine because your virtual server is actually running in a cluster on several machines. This also minimizes the danger of outages due to upgrades or individual machine failure. You save money because you pay a base rate for the virtual server, and then only pay for the computing resources you actually use. This could seem to get expensive depending on rates, but if you're using more computing power it's likely because you have more customers or sales leads, and this is a good thing. In fact, virtualized servers start for much less than leased physical servers.
In short, virtualization can save you money, increase reliability, and allow you to do more, no matter what size your business is.
