Yes it's finally here... you can now model your mailbox servers for a DAG spread across 2 datacenters in Active/Active mode.
Released: v12.3 of the Exchange 2010 Mailbox Server Role Requirements Calculator
http://msexchangeteam.com/archive/2010/09/17/456327.aspx
Direct Link to download the Calculator: http://msexchangeteam.com/files/12/attachments/entry453145.aspx
Below an extract of the tracking log found here: http://msexchangeteam.com/archive/2010/01/22/453859.aspx
Tracklog
==Version 12.3 Updates==
Bug Fixes
- Fixed Total Number of Databases / Server calculation to deal with scenario where lagged copies are deployed in both datacenters for Active/Active (Single DAG) scenario
==Version 12.1 Updates==
Bug Fixes
- Optimized Number of Active Databases after First PDC Server Formula removing redundant bad code and enabling single database scenario.
- Fixed Number of Required Mailbox Processor Cores for both PDC and SDC calculations to take into account the situation where the required megacycles to support the active load is less than the number of megacycles per core.
- Optimized Number of Required Mailbox Processor Cores for both PDC and SDC calculations to not assume all required cores would be 100% utilized by changing how rounding works in the formula.
- Fixed Number of Active Databases / PDC Server (After Second PDC Server Failure) formula to only report a value for the following scenarios: HA Only 3+ HA copies, 4+ servers; HA, Site Resilience, No activation block, 3+ total HA copies, 4+ total servers; HA, Site Resilience, activation block, 3+ PDC HA copies, 4+ PDC servers.
- Fixed standalone scenario to expose the total number of databases being deployed in the results section when there are multiple servers.
- Fixed Number of Active database calculations (after first PDC server failure) to take into account activation block.
- Fixed the Number of active databases during normal runtime formula to round up.
- Various formatting and text cleanup across all sections.
- Fixed data error validation statement for IOPS and Megacycle Multiplication Factors to specify the supported value must be 1 or greater.
Enhancements
-
- Open a web browser and got to www.spec.org.
- Click on Results, highlight CPU2006 and then select Search CPU2006 Results.
- Under Available Configurations, select SPECint2006 Rates and click Go. Under Simple Request, enter the search criteria (e.g. Processor matches x5550).
- Find the server and processor you are planning to deploy and take note of the result value. For example, let's say you are deploying a Dell PowerEdge M710 8-core server with Intel x5550 2.67GHz processors (2670 Hertz); the SPECint_rate2006 results value is 240.
- Added Megacycle Multiplication Factor – this works exactly like the IOPS Multiplication Feature does and was added as a result of RIM providing E2010 guidance on megacycle impact due to Blackberry devices.
- Active/Active user distribution scenarios. Yes, really! An Active/Active user distribution architecture has the user population dispersed across both datacenters (usually evenly) with each datacenter being the primary datacenter for its specific user population. In the event of a failure, the user population can be activated in the secondary datacenter (either via cross-datacenter single database *over or via full datacenter activation).
- Added a new worksheet/section that documents the Activation Scenarios for DAG deployments.
- Added error reporting validation logic if HA solution results in greater than 16 servers in a DAG to not show any results, since the design is invalid.
- Dumpster size calculations have been optimized as calendar versioning storage has been reduced from 5.8% impact to 3% impact in SP1.
Posted
09-18-2010 7:31
by
Tonino