Windows Web Hosting, Web Technologies, etc
Archive for August 8, 2009
Elastic Computing AKA Cloud Computing as provided by Applied Innovations
Aug 8th
I broke this post up into a series of posts. The first post (just published) gives an overview of ARR in IIS7 and why it’s cool. This post is going to talk about the first stage of cloud computing we’ve deployed at Applied Innovations and the benefit’s of it such that you could use it today to control your own hosted IT Infrastructure costs.
About the Applied Innovations Cloud Platform.
Applied Innovations is a charter member of Microsoft’s Dynamic Datacenter Alliance. The alliance is composed of a handful of hosters, ISVs and system integrators that have deployed a solution on Microsoft’s Dynamic Datacenter toolkit. Our implementation of the toolkit is a highly available cluster of VPS servers. These VPS servers or a series of physical servers that use a high speed, redundant SAN cluster for storage and in the event of a hardware failure or need for maintenance the virtual machines running on one of those servers will automatically fail over to one of the other nodes to keep the services and applications running 100% of the time.
In addition, our own solution also affords the user the ability to achieve a level of cloud computing called: elastic computing (a hot topic these days as Cloud computing is all the rage). Elastic computing is the ability to dynamically scale up and down your computing resources as your needs change. For instance, let’s consider the case of dedicated hosting.
When the Dedicated Server Meets the Cloud.
In the past when a hosting customer outgrew shared hosting they needed to move to a dedicated server. Moving an advanced web application (such as a live ecommerce website) from one server to another is time consuming and often costly so it’s something most store owners don’t want to do frequently. With that in mind they would often look at dedicated servers and have to predict their need for hardware not only today but often 6, 12 or even 24 months into the future or they could find themselves moving again in the near future. The problem with this is that often they found themselves purchasing more hardware and taking on a larger cost today in preparation for tomorrow than they needed and sometimes could afford.
Further, today we all know just how quickly the economy can turn and how fast you can find that your expected growth can disappear and all of a sudden find yourself needing to scale down instead of up. Or perhaps your business is seasonal and you find that you really only need a full dedicated server a few months out of the year and don’t need to pay for the full solution the other 9 months of the year.
At Applied Innovations, Our Dynamic Datacenter Solution is the answer to these very issues. With one of our highly available clustered Windows virtual dedicated servers (VDS servers) you’re able to scale up your diskspace, your memory, your processing power and likewise scale down your diskspace, your memory and even your processing power with just a click of a few buttons. This puts you in control of your hosted IT infrastructure and allows you to adjust your costs in line with your business.
What about Elastic Scalability? I thought that was Cloud?
The other hot topic in Cloud is Elastic Scalability. This is where your web infrastructure is hosted on multiple virtual machines (which may be elastic computing VMs) and is load balanced across these nodes. you then have the ability to expand and contract the number of nodes your site runs on as your traffic and business demands change.
We believe this too is a vital component to cloud computing but for 80% of the hosting business on the market today, it’s just not a necessity.. today. For that upper echelon of hosting customers (the Amazon, the Twitter, the Facebook type company) this is a necessity but for most companies this level of cloud computing is just an extra expense and in today’s economy extra expenses are not needed.
What’s next in the Cloud?
This is only the first phase of our own Cloud Initiative and there are of course other notions of cloud besides elastic computing and elastic scalablity that I didn’t hit on. What I can guarantee is our team is working as diligently as ever to continue to work on our Cloud Initiative and will continue to evolve our service offerings. In the meantime if you want to cut your dedicated hosting costs and have the ability to adjust your hosted IT infrastructure costs as your business adjusts, contact our team about our Managed Windows VPS Hosting solutions built on top of the first stage of our Cloud infrastructure.
A little info about Application Request Routing (or ARR) in IIS7.
Aug 8th
In the 10 years I’ve been working with Windows as a web hosting platform (and the 15 or so I’ve been working with web servers period), I can confidently say that IIS7 is without question the best web server platform to date. Microsoft has been working on a plug-in for IIS called Application Request Routing (or ARR). The first version of ARR provided a suite of tools that allowed it to function as a Application Layer Load Balancer! The recently updated beta version extends that functionality and now makes it easy to turn a Windows Server running IIS7 into not only a Load Balancer but also a Reverse Proxy / Cache Server. Meaning you could potentially power a CDN using a bunch of Windows Servers.
What’s an Application Layer Load Balancer?
First I should define a load balancer. When you have a very popular website (think yahoo, Amazon, etc) you need a bunch of web servers to accept all those web requests. One server just won’t cut it. In order to keep things all tidy and in order you use a machine that controls those requests and which web server it goes to. You extend that machine and have it start monitoring your web servers and your web visitors and it can start managing those requests and directing it based on things like:
- The visitor’s browser type
- The visitor’s IP address
- The type of content being called (JPEG, MOV, HTML, ASPX, etc)
Then you get a step fancier and not only direct those requests based on the request but also on the state of the server, who’s had the most bandwidth, which server do you prefer requests to go to, which server is working hardest? etc.
Now the Layer 7 Load Balancer or Application Level Load Balancer can distribute requests based on the various application specific information. This could be a cookie for instance or information within your app itself.
NOTE: I’m not going to go into this any further you have the basic understanding of a load balancer and I’m not going to deal with NLB and layer 4 and all that fun stuff. Go Bing! it (yeah I said Bing!).
Wait aren’t load balancers expensive, complicated and require a whole other level of geek?
You bet! But they’ve gotten better and there’s a fair number of open source ones out there (in fact, I was reading earlier that wordpress.com was running on an open source one).
Well didn’t Windows Always have this?
Not really. Windows had Network Load Balancing which would allow you to run a couple servers with the same application on it. It would monitor ping on the machines and if one machine died all requests went to the other. So that was really load balancing at the machine level. This was good enough for most but what if the machine was fine but your service (perhaps IIS itself crashed?) . Then they had Clustering services. Now clustering services was service aware. Meaning if your service (perhaps IIS ) crashed on one of the machines it would failover to the other machines but it didn’t really know anything about your website’s application (perhaps an ecommerce store) so you could have a server online and the web service online but the actual application on that server failed and still be sending requests to it. So in comes those clever folks in the IIS team at Microsoft with Application Request Routing (ARR … Matey!).
Video Introduction to Application Request Routing and more info
Microsoft has a great video on ARR and it’s geared towards webhosts but does a fantastic job of showing the functionality of ARR and just how easy it is to configure and make use of. You can see that video on the IIS.net site at:
http://www.iis.net/extensions/ApplicationRequestRouting
In addition to the video they really spell out all of the features and benefits to ARR in detail. Some of those features include:
- HTTP based routing decisions.
- The ability to offload SSL requests
- The ability to offload compression
- Health monitoring of live traffic and specific URLs to monitor and adjust routing of requests.
- Support for extensive debugging through Failed Request Tracing Rules.