[Okta Webinar] Learn how to a build a cloud-first strategyRegister Now

  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 815
  • Last Modified:

Application Aware NLB and Clustering solutions

Hey all,

I'm doing a survey in order to find the best way to assure the lowest downtime and the best load management to the critical applications in my organization
I'm using Windows Server on VMware VI 3.5.4(Upgrading to vSphere soon)

Right now we're using MSCS to provide Application aware clustering (even for single node Clusters). On the other hand, we are using Microsoft NLB for load balancing.  

I'm looking for a way to combine the two things, and to lessen the complexity caused by using MSCS in a VMware environment.

 The applications are all based on .Net, some are web applications, and the others are client-server.

Any Ideas?
2 Solutions
wow man you got yourself quite a thing there :)

try making it easy not complex.
every IT structure is going to fail this is %100
there is not IT thing or application which is %100 stable.

so having this in minde make your setup as easy as possible so you can restore it fast.

wrdanaAuthor Commented:
I'm a woman :)
I know nothing is 100% safe, but there are solutions out there that can provide what I'm looking for.
For example - Oracle RAC Architecture, which only applies to Oracle DB.
Any load balancing appliance should be able to do the trick -- they'll all monitor the TCP/UDP packet flow of your applications and should be able to make traffic management decisions based on that.  They should all (with the exception of things like round-robin DNS) be able to figure out when a server has failed and alert you while sending the traffic to other servers.

As far as 100% uptime, that's really the goal of the application availability/ traffic management industry.  Yes, artoaperjan is right that it'll never be quite 100%, but certainly you can take certain precautions to make sure that you get as close as possible.

A clustering solution is certainly the first approach -- that way, if a server fails you have other servers to handle the load.  Also make sure that whatever clustering solution you select, it has the ability for failover in case the solution itself fails.

I'm familiar with Coyote Point Systems load balancers.  These are usually deployed in pairs so that if one system fails, you are notified via email/SNMP/logging/etc, and the other system takes over for it so your applications remain available.  This is how all of the main players in the traffic management market work as well.

The thing that really sets different companies apart is their products' knowledge of particular applications, and the configuration options that you have to set up how failure is detected.

As far as Coyote Point, the Equalizer load balancer has an add-on called VLB which allows integration with VMware Infrastructure (or vSphere) to perform either scripted or trigger-based rules such as spinning up a new server if one of the active servers fails, or rebooting the failed server.  (I bring this up since you said you're using VMware, and as far as I know Coyote is the only company that offers such a thing).

To summarize, when selecting a clustering/High Availability solution, you should look for:
1) redundancy both within the application monitoring, and within the load balancer itself.
2) different methods of determining your application's status (from Layer 3/4 -- ICMP probes to Layer 7 -- sending specific requests to your application)
3) different methods of notification if a problem is found (such as SNMP, email, flashing lights, logging, etc).
4) probably most importantly, sit down and take a good look at your applications and figure out which ones need high availability -- and then specifically discuss those applications with the load balancer sales people when you talk with them to make sure that it will be supported in a way that you require.

Hope this helps!

Featured Post

Upgrade your Question Security!

Add Premium security features to your question to ensure its privacy or anonymity. Learn more about your ability to control Question Security today.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now