Number of Mailboxes

Total number of mailboxes that shall be deployed at a site. Based on the size of the environment and the profile of the users, the sizing tool will determine the appropriate number of servers and DB instances per server.

Size of Mailboxes

For storage capacity estimation, this is the mailbox size limit.

Average message size

By default, the average message size is 50KB.

User profile

Pull-down menu with the options of Light, Average, Heavy, and Very Heavy. The User Profile provides an estimate of the amount of email sent or received per day.

Mailboxes moved per week (%)

Mailbox moves are part of the administration of Exchange 2007 and can occur regularly in order to tune the system for optimum performance. Users consume resources differently and it may be necessary to move different user mailboxes to different DB instances or possibly different email servers accordingly. Mailbox moves won’t necessarily effect the overall configuration of the mailbox servers, it will effect Log sizes as log files are created for the move.

Percent of cached mode users

This is the percentage of users operating in cached-mode versus online-mode. Cached-mode users run from a cached instance of their mailbox in their email client. Online-mode users access their mailbox directly from the mailbox servers and impose more IOPS than a cached-mode user. This number is used to determine Total IOPS.

Recovery Storage Group

A recovery storage group is used as part of the disaster recovery plans to stage a DB instance. Enough capacity needs to be available to handle the largest of the databases that will be restored on that server.

Clustering

Pull-down with the options of, Cluster Continuous Replication (CCR), Local Continuous Replications (LCR), Standby Continuous Replication (SCR), Single Copy Cluster (SCC) or none.
CCR does log shipping from the primary mailbox servers to standby servers and are then replayed to disk storage on the secondary system.
LCR replays the log files to additional disk on the same primary mailbox server. LCR storage can be on the same array or a completely unique disk system attached to the same server. The mailbox server in LCR could be a single point of failure (SPOF).
SCR does log shipping from the primary mailbox servers to other (possibly remote) server(s) and are then replayed to disk storage on the other server(s). Note: SCR does not provide automate failover!
SCC uses a single instance of the database and log files shared between clustered servers and depends on additional clustering software such as Microsoft Cluster Server (MSCS). The disk data in SCC could be a single point of failure (SPOF).

It is important to know what Clustering/Replication options are considered not only for storage and server counts, but I/O profiles vary depending on the type chosen.

If SCR with or without CCR Clustering, How many Targets? [0]

SCR clustering can be in conjunction with CCR. Also, multiple SCR targets can be created for higher availability. Use this field to select the number of SCR targets if SCR is to be used (regardless of selection in Clustering field).

Indexing

Pull-down menu with the options of Enabled or Disabled. If content indexing is enabled it requires an additional 5 percent capacity per storage group.

Data Growth (%)

Planned growth factor.

Backups

Pull-down menu with the options of Full, Daily Incremental Weekly Full and Daily Differential Weekly Full. Depending on the backup strategy used, the number of log files kept will either be set to three days (Full) or 10 days for incremental or differential options. This will determine how long log files need to be retained and is used to calculate the Log LUN size.

LUN Overhead (%)

Percentage of free space calculated into LUN size.

Dumpster Days

By default in Exchange 2007, deleted items are retained in the database for 14 days before they are permanently removed during a maintenance activity. The space on disk is not calculated into the mailbox quota and therefore consumes additional space and will be used to calculate the actual space per mailbox user on disk.

Percent of Simultaneous Users (%)

Users with active mailbox connections which have active server threads. This is typically between 75-80 percent of the total mailboxes in a production environment. ActiveSync and Black Berry devices can affect this number as they also create an active connection.

IOPS Overhead

By default in Exchange 2007, I/O Overhead is set at 20 percent to ensure adequate headroom and allow for abnormal spikes in activity that may occur from time to time.

Number of Messages (avg) Received per Second from the Internet

By default, this is set to 20 messages.

Maximum Size Message Allowed

By default, the maximum message size allowed is set to 15MB.

Include Recovery Group Size in Log Capacity?

Pull down with Yes or No as the options. By default, this should be set to Yes.

Type of Drive

Pull down with the following options: FC 146GB 15K, FC 300GB 15K, FC 73GB 15K, FC 146GB 10K, FC 300GB 10K and FC 73GB 10K. Once the sizing tool calculates the size on disk that will be required for the DB LUNs and the Transaction Log LUNs, changing this variable along with RAID Protection (below) will describe the number of disks that shall be required.

RAID 10 or RAID 5

RAID 10 is Stripping and Mirroring
RAID 5 is Striped Data Drives with Distributed Parity

Pull down on with the following options: RAID10 or RAID5 this will determine the number of disks required based on the RAID protection selected. Note that although RAID 5 incurs a write hit overall in some configurations it may require fewer drives.

If Raid 5 What Number of Data Drives? N+1

How many data drives per Raid 5 stripe. Set to 7 by default.

Spare Drives

Number of spare drives per array to be factored into number of drives needed per array.

Dedicated Client Access Server

For smaller configurations a Dedicated Client Access Server may not be necessary. Please refer to sizing guidelines in Reference Section for Microsoft’s recommendations on Client Access Server configurations.

Array Selection

Pull-down with the following selections for the array to configure: 2510, 2530, 2540, 6140 Small, 6140, 6540, 9985V and 9990V. Choose the array to be configured.

Separate Array for Cluster CCR/LCR

If space permits in the maximum configuration of the array selected in ‘Array Selection’ based on ‘Type of Drive’, ‘RAID 1 or RAID 5’ and ‘If RAID 5 What Number of Data Drives’ , it is possible to locate CCR or LCR replicas within the same array. This may be desirable for a less complex solution but may impose additional single points of failure into the configuration.