cancel
Showing results for 
Search instead for 
Did you mean: 

Information on the CPU/Memory overhead of using SFWHA-DR versus native Windows

ChipApple
Level 3
Partner Accredited Certified

Hi everyone,

We are currenlty having a bake-off for Exchange 2010 of using SFWHA 6.0.1 or native a native Microsoft DAG for a COOP/DR solution at he customer site I support. 

My question for Symatec would be if there is any documentation on the amount of extra overhead on each Exchange 2010 mailbox server when using SFWHA versus just the Windows 2008 operating system itself?

The plans are for:

6 Mail Servers
4 Databases per server
Windows 2008 R2 with each virtual server with 64 GB RAM

If there ae any planning guides from Symantec for Exchange 2010 using SFWHA that would be useful as well.

Thanks in advance,
Chip

 

1 ACCEPTED SOLUTION

Accepted Solutions

D_White1
Level 4
Employee Accredited

Hi Chip,

Interesting that you assume there will be additional overhead when using SFWHA!  8o)

SFWHA 6.0.1 includes the IMF (Intelligent Monitoring Framework) for several VCS agents which actually reduces the CPU overhead for monitoring cluster resources.  Essentially, this allows you to deploy an event-driven monitoring framework that removes the requirement to poll every resource every 60-seconds, therefore reducing the overhead on the system for the HA software.

In terms of the overall solution performance metrics, this would require more in-depth analysis around the replication methods used, the solution design etc. before anybody could realistically comment on any performance aspects.

Chapter 5 of the EXchange 2010 Solutions Guide contains information on planning your deployment:

https://d1mj3xqaoh14j0.cloudfront.net/public/documents/sfha/6.0.1/windows/productguides/pdf/SFW_HA_D...

Dave 

View solution in original post

2 REPLIES 2

D_White1
Level 4
Employee Accredited

Hi Chip,

Interesting that you assume there will be additional overhead when using SFWHA!  8o)

SFWHA 6.0.1 includes the IMF (Intelligent Monitoring Framework) for several VCS agents which actually reduces the CPU overhead for monitoring cluster resources.  Essentially, this allows you to deploy an event-driven monitoring framework that removes the requirement to poll every resource every 60-seconds, therefore reducing the overhead on the system for the HA software.

In terms of the overall solution performance metrics, this would require more in-depth analysis around the replication methods used, the solution design etc. before anybody could realistically comment on any performance aspects.

Chapter 5 of the EXchange 2010 Solutions Guide contains information on planning your deployment:

https://d1mj3xqaoh14j0.cloudfront.net/public/documents/sfha/6.0.1/windows/productguides/pdf/SFW_HA_D...

Dave 

ChipApple
Level 3
Partner Accredited Certified

Thanks Dave,

I will check out the document.

Best Regards,
Chip