Exchange 2007 Mailbox Server Role Storage Requirements Calculator updated to v 13.0

[The_Exchange_Team]

It's been a while since we announced an update to Exchange 2007 Mailbox Server Role Storage Requirements Calculator.

Yesterday we have updated the calculator to version 13.0. Updates:

• We have updated the log generation numbers per message profile to be in line with our updated guidance.
• In v11.8, we decided to list the database cache per mailbox in the Storage Requirements results section.  However this lead to confusion because it was named "Memory Profile / Mailbox" which implied that you would only utilize the associated amount of memory with the message profile (i.e. 5MB with Heavy profile), however that is not always the case. For example, 1200 2GB Light message profile mailboxes only requires 5GB of RAM (1200 * 2MB + 2GB), however the design requires 13 databases, which with SP1 requires 6GB of RAM.  ESE will utilize 4GB of that RAM for the cache.  As a result, 4096MB / 1200 ≈ 3.5MB per mailbox. So to make this clear, we have changed this text to be "Database Cache / Mailbox" which indicates how much cache is available per mailbox.
• In the scenario where you override the IOPS prediction formula for your mailbox tiers, we have adjusted the "Read:Write Ratio" input to allow you to enter any read percentage you would like, rather than restricting you to a few key ratios.
• We updated the "Database Reads / Mailbox" calculation description.
• We updated the Log Replication Requirements worksheet, simplifying the data displayed in the results section.
• We have included new functionality for log replication requirements.  You now have the input options for entering your network link type and its associated latency.  These options are then used to recommend TCP/IP optimization settings for Windows Server 2003 when utilizing geographically dispersed clustering and/or standby continuous replication.  In addition, if the chosen network link cannot sustain the throughput requirements for log replication, we will recommend an appropriately sized network link and Windows Server 2003 TCP/IP optimization settings.

For full list of updates (per version), go here.

For the explanation of different tabs and how the calculator works, go here. Yup, we updated that too!

Finally, to get the new calculator - go here.

- Ross Smith IV

[#2]

Great update, thank you! Any thoughts to giving us a dropdown options to select user profiles which send more mail than they receive? Thanks again.

[#3]

bday: Suppose you have 35sent/15received for the average user (which is highly, highly unlikely, unless most of your "users" are spam-bots). This would equate to the 10sent/40received dropdown selection, since the transaction log volume is dictated by the average of 50 messages/day, and not a ratio of sent-to-received. If you're skeptical of your message counts like I am, run the Exchange profile analyzer (EPA) to determine daily sent/received average per user; EPA will also provide you with the average message size. If you derived your ratio using a freeware message-tracking-log-parsing tool, don't be surprised to find inaccuracies. I've seen some free 3rd party tools grossly inflate the # of sent items because they aren't intelligent enough to ignore the same message that listed multiple times in the tracking log. EPA doesn't rely on tracking logs; rather, it will login and scan messages in each mailbox to generate daily statistics (though this might be a little skewed if you choose to scan an older timeframe, since users tend to delete more items over time).

[#4]

Hi there. I used EPA to gather all of my org's figures for the E2K7 calculator (allong with the collectlogs.vbs script to get our hourly E2K3 transaction log rate). I could easily be interpreting its figures incorrectly. Such as should the metrics gathered for Sent Items, Forwarded Items, and Replied Items all be tallied up and read as one cumulative "Sent" total or is the sent items figure itself the actual total and replied/forwarded just extra fluff figures if one was curious about their environment?

I've not been able to find clear documentation on this and haven't gotten an answer in the past when posing the question. Thank you.

[#5]

Bday - Use Sent Items + Deleted Items (or similarly worded since I'm going from memory) and Received Items + Deleted Items metrics for determing your send/receive rate.  Also, I would limit your collection to 1 business week or less to ensure a more accurate data sampling.

Ross

[#6]

Sorry, second time i post this...

does anybody know if i still can download exchange 2k7 sp1 beta 2. Cause otherwise i can`t upgrade from sp1 beta 1 to rtm..

[#7]

Great Tool,  Where can I find one for sizing the other roles - Edge, Hub, UM and CAS?

[#8]

Thank you, Ross, I'll go back and redo the numbers using simply those metrics.

[#9]

I've been experimenting with this tool, and I can't get it to behave the way I believe it should.  For example, I specify 100,000 users and 'Yes' to 'Use Recommended Maximum Database Sizes?', but it results in 49 databases, each around 1.1 TB in size. Using a smaller mailbox count - like 30,000 results in 49 databases, each around 303 GB.  Does anybody know why I would get these results?  Is the calculator working as expected or as it is designed to do?  I downloaded v 13 for this exercise.  Thanks.

[#10]

Hey Sean,

Yes the calculator is working correctly.  The calculator assumes that all users entered will be placed on a single server.  If you want a smaller number of users per server you will have to enter that number instead.

[#11]

Ross,

Thank you for your reply.  I'll see if I can find the maximum number of users I can fit on a server while keeping the target database size at Microsoft-recommended levels.  That should solve my problem.

[#12]

We're finding this tool very useful.  Thank you for producing it.  I read through the documentation which explains each input and output in the calculator, and I am wondering why the Restore LUN Size / Node is sized to support "up to 7 databases and 7 transaction log sets".  Normally, we only ever need enough to support one database and transaction log set with Exchange 2003.  Is there something different which needs to be considered regarding Exchange 2007 or is this a Microsoft recommendation?  If so, can you direct me to that article which talks about it?  Thank you.

[#13]

Hi Sean,

When sizing the restore LUN, we essentially use the following rules:

If LUN Architecture = 2 LUNs / SG
- We size the restore LUN using ((DB_Size_With_Overhead + Log_Size_With_Overhead)) * (IF(#_SG<=7, 1, #_SG/7))

where #_SG = number of storage groups

Since you have up to 50 storage groups, there is a risk that multiple SG failures could occur, so we want to size the restore LUN to ensure that you can survive multiple failures.

If LUN Architecture = 2 LUNs / Backup Set
- We size the restore LUN using ((DB_Size_With_Overhead + Log_Size_With_Overhead) * #_SG_In_BackupSet)

where #_SG_in_BackupSet = number of storage groups contained within a single LUN

Basically with this formula we want to ensure you can restore all the storage groups should use lose an entire LUN set.

Hope that helps,

Ross

[#14]

Great tool guys!  Thank you!  I am having trouble though, as I do not see a place to specify the assumption for data growth over time.  For example, my scenario calls for ~3 TB of online disk storage split up on a total of 6 LUN’s... Based on this, how long will my storage last me?  I do not see where we can set the time frame that the recommended storage requirements will scale to based that the continuation of the trend that we specify to derive the storage / LUN requirements.

Please explain thanks.

[#15]

Ross,

Thanks for the explanation regarding the sizing of the RSG.  When you state that with up to 50 SGs, "there is a risk that multiple SG failures could occur, so we want to size the restore LUN to ensure that you can survive multiple failures".  

Generally, we treat the RSG as the mechanism by which we restore a mailbox or items to a mailbox or export a mailbox to a PST, but you seem to be saying that the RSG is used for disaster recovery purposes.  Any time we've had a disk subsystem failure that caused stores to go offline, we've never used the RSG to recover.

I must be missing something or haven't encountered the kinds of emergencies that you have.  I just can't seem to understand the need for a large RSG/maintenance volume.  Can you describe a scenario where multiple SGs failed and you resorted to utilizing the RSG disk space?

Again, thanks.  I appreciate all of your feedback and assistance, and the tool is a real time saver for us.

Regards,
Sean

[#16]

Hi Jesse,

Yes you are correct in that the tool does not have an option to define length of time and growth over that time.  For now you have two options:

1.  Add in the additional planned mailboxes up front.
2.  Utilize the data and I/O overhead factors.

I will mark this down in one of my to-do's and see if I can get it into a future version of the product (no promises!).

Ross

[#17]

Hi Sean,

One example would be implementing a dial-tone scenario.  

Ross

[#18]

Ross,

Thank you for adding the input factor, Projected Mailbox Number Growth.  This will really help to add growth projections to the model.

Jesse Kegley

[#19]

Question,
how do I convert LUN to physical drives?

We are not using a SAN but a Direct connect device.

I am assuming that the I will have spindles for the OS
Spindles for the database
spindles for the logs

and depending on the type of replication service spindles for the replicaitons. Am I on target or am I way off base?

thank in advance
Dmcland

[#20]

Hi dmcland,

A LUN doesn't necessarily correspond to a physical set of disks (aka RAID group).  It could simply be a partition on the RAID group or it could be the RAID group itself.

So the calculator will tell you how many LUNs you need and the host I/O requirements you need to satisfy.  Based on that information you can determine how many spindles you need to satisfy both capacity and I/O requirements (remember to take into account RAID penalty).  Then you can simply carve out the LUNs the way you want.

But yes, basically you will want spindles for the OS/Applications (RAID-1 set), spindles for the database (RAID-1/0 or RAID-5 sets), and spindles for the logs (RAID-1/0 set).  And then of course additional spindles for the continous replication copies.

If you want some additional information on determining number of spindles, please see http://technet.microsoft.com/en-us/library/bb125079(EXCHG.65).aspx.

Ross

[#21]

Ross,
thank you,
one final question

SP1
exchange servers 1
CCR

No dedicated maintenance / restore lun? Why?

14
20%
2%

900 users

%
sent 20/80 received
50 K
250 Quota

IOPS .16


I get
341 G for database
182 Logs

What is database LUN Disk Space required? 916G????

I plan on direct connect storage

3 database
300 users per database

What else should I check

Darrin