Table of content
- Mailbox Configuration
- Backup Configuration
- Storage Configuration
- Processor Configuration
- Log Replication Configuration
- Environmental Customization
- Where to Go from Here
In Part 2, we continue to enter values into the official Microsoft sizing tool – Exchange Server Role Requirements Calculator. In this document, we cover mailbox, backup, storage, processor, and log replication configuration. This document also describes environmental customization techniques. The current version of the calculator has support for Exchange Server 2016. All of the recommendations are the same as those for Exchange Server 2013, except for processor recommendations which have increased slightly.
You can download the tool from TechNet
Input for the mailbox configuration can be generated using various tools, such as:
- Email statistics scripts
- Environmental report and log parsers
- Other business and technical requirements
The tool is set up so you can enter user mailbox configuration values based on tiers. For example:
- Tier 1 could be a normal user. The following figure shows fields for tier 1.
- Tier 2 could be heavy users.
- Tier 3 could be VIP users.
- Tier 4 could be for contractors and mobile device users.
Depending on the user tiers, values must be entered into the calculator. Some of the key values to be entered are:
- Total number of users.
- Projected mailbox growth percentage.
- Message statistics from Exchange profile analyzer (average message size and total
- send/receive e-mails per mailbox per day).
- Mailbox size with limits.
- Personal archive mailbox size.
You can accept the default values for the other entries, except IOPS Multiplication Factor and Megacycles Multiplication Factor. The Exchange Server Role Requirements Calculator predicts the Input/Output operations per second (IOPS) requirements for each user. Many users utilize mobile devices such as iPhone and mobile device management (MDM), which require high CPU and IOPs utilization, is often in use. Consider the following when designing the IOPS and CPU Megacycles multiplication factor for mobile users.
- IOPS Multiplication Factor: 2.0 (increase this requirement by 100% for each user).
- Megacycle Multiplication Factor: 1.5 (increase this requirement by 50% for each user).
Depending on requirements and infrastructure, you can conduct fewer backup operations using Exchange Native Data Protection or other backup solutions. Exchange Native Data Protection uses circular logging to purge transaction logs and provides single-item recovery for data resiliency. Implementing Exchange Native Data Protection requires at least three non-lagged copies of the database. To protect against logical corruption, it is recommended to have at least one lagged database copy.
For business continuity and disaster recovery, you may want your backup methodology to include both Exchange Native Data Protection and Volume Shadow Copy Service (VSS) backups for long-term retention.
Backup/Truncation Failure Tolerance allows you to specify the number of days that logs are retained when no backup is run and logs are not getting purged. We recommend you set this value between 3 and 5 to retain sufficient free disk space for retaining logs for the specified number of days without performing backups.
Exchange Server storage sizing includes estimating the storage space to accommodate databases, content indexes, transaction logs, growth factors, and unexpected overhead growth.
Storage sizing also involves calculating the IOPS required for the databases, content indexes, and transaction logs. These values correspond to performance driving factors for users. Better IOPS improves the user’s client performance.
The storage configuration settings also require you to determine disk capacity, type, and speed. Once these settings are determined, you can enter the appropriate storage disk options for two datacenters in the Datacenter 1 Server Disk Configuration and Datacenter 2 Server Disk Configuration areas of the calculator.
A key factor during the design process is determining the CPU/processor requirement. CPU sizing is done in terms of Megacycles, and is determined for each mailbox associated with a user profile. In Exchange 2013, Megacycles utilization is on the high side for active copies and on the low side for passive copies. It also varies when the server is installed with a multi-role setup. The Processor requirement for Exchange Server 2016 is the only difference from Exchange Server 2013. See this Exchange Team Blog post for details.
The SPECint_rate2006 value is a required value that must be specified. To ascertain this value, use Microsoft’s Exchange Process Query Tool, which you can download from http://www.spec.org/. Enter the process model number into this tool and the tool replies with the average SPECint_rate2006 value based on the number of cores selected and hardware vendor.
The right column of the tool allows you to determine the exact SPECint_rate2006 value for servers from various vendors (see Figure 3). The tool also helps to determine the SPECint_rate2006 value based on the number of CPU cores allocated to a mailbox server configured for a virtual environment.
After you have all the necessary information, update the number of cores/servers and the SPECint_rate2006 value based on the Exchange Process Query Tool. If the mailbox server is configured for a virtual environment, add 10% as the Hypervisor CPU Adjustment Factor.
Log Replication Configuration
This configuration is necessary only when you have to replicate database copies between two datacenters. This configuration estimates the network bandwidth requirement between the two datacenters for log replication for given DAG databases.
Enter the percentage of transaction logs generated for each hour of the day. You can collect this information from the current environment using the information outlined in this blog post. Enter the information gathers about logs into the Exchange Role Requirement Calculator.
The Network Configuration fields allow you to update the available network link type (speed) and network link latency between the datacenters.
This is the last configuration in the input sheet of the Exchange Role Requirement Calculator. Update the DAG node names and the DAG database base name (the base name is the starting part of the database name and must be unique in the organization). This information is used to generate scripts and configuration files, which create and configure the DAG and its databases with database copies.
Where to Go from Here
You have now updated all the necessary information in the Calculator. The Calculator will generate multiple result panes/sheets containing design and configuration recommendations. Part 3 and Part 4 of this article series discusses these result sheets.