Load Balancers and SEO
For years now I hear people say you cannot use a load balancer and expect to see SEO traffic. In some cases this can be true it all depends on how you choose your equipment and configure it.
There are some basics that you need to understand, and not one brand is the total solution to your problems. First the number one thing to look at is: Does it alter the URL’s in any way?
If it changes the URL’s from, www to say www1,www2,www3 or something similar you need to be careful. Yes this still can work, but you need to place logic in the system to look for spiders. If a spider shows up do not ever let me go to any other node than www, most search engines will consider it duplicate content to have the same data appearing on so many load balanced nodes. Thus it’s critical that this be addressed in advance.
Additionally you may have problems when someone is trying to link to your site down the line. For example if you have a link to www10.yoursite.com/cool-widget-2.html, now you are faced with someone linking to a load balanced node. Thus if a spider where to try and follow the link, it could pass credit to the wrong page. Thus it is critical to have logic that will make sure that the search engine will see only the www version of the site.
You can accomplish this by a simple 301 logic-set, but make sure it hits the exact same page on the main node.
The best possible solution for load balancing is what I call a “smart” environment or might be called round robin. Where the URL is never changed and this can be accomplished by a rotating DNS logic. Which will use a certain set of IP’s, however never showing a different URL, a process which can easily be accomplished.
There are a few fundamental problems with the Round Robin approach; it can be very difficult to be sure that your load is equally being distributed among the boxes. Cached IP’s in a browser and certain ISP’s may send it to the wrong node and kill your site. However there are a few hardware manufacturers such as Cisco who can handle these issues.
You might also want to consider using a service such as Akami, whom can deliver your content from a single source without overloading your servers. I spoke with Akami last year and was given a tour of how their SEO friendly technology works. Seems like some cool stuff, but not cheap!
There are some basics that you need to understand, and not one brand is the total solution to your problems. First the number one thing to look at is: Does it alter the URL’s in any way?
If it changes the URL’s from, www to say www1,www2,www3 or something similar you need to be careful. Yes this still can work, but you need to place logic in the system to look for spiders. If a spider shows up do not ever let me go to any other node than www, most search engines will consider it duplicate content to have the same data appearing on so many load balanced nodes. Thus it’s critical that this be addressed in advance.
Additionally you may have problems when someone is trying to link to your site down the line. For example if you have a link to www10.yoursite.com/cool-widget-2.html, now you are faced with someone linking to a load balanced node. Thus if a spider where to try and follow the link, it could pass credit to the wrong page. Thus it is critical to have logic that will make sure that the search engine will see only the www version of the site.
You can accomplish this by a simple 301 logic-set, but make sure it hits the exact same page on the main node.
The best possible solution for load balancing is what I call a “smart” environment or might be called round robin. Where the URL is never changed and this can be accomplished by a rotating DNS logic. Which will use a certain set of IP’s, however never showing a different URL, a process which can easily be accomplished.
There are a few fundamental problems with the Round Robin approach; it can be very difficult to be sure that your load is equally being distributed among the boxes. Cached IP’s in a browser and certain ISP’s may send it to the wrong node and kill your site. However there are a few hardware manufacturers such as Cisco who can handle these issues.
You might also want to consider using a service such as Akami, whom can deliver your content from a single source without overloading your servers. I spoke with Akami last year and was given a tour of how their SEO friendly technology works. Seems like some cool stuff, but not cheap!
Labels: load balancers, seo
by Aaron Shear
0 Comments:
Post a Comment
<< Home