One final key to providing continuous availability for your Exchange deployment is the ability to set some guidelines for determining whether the system is meeting the organization’s business objectives for implementing the system in the first place. E-mail has become a key mission-critical application within most organizations, but just how much service degradation can the organization tolerate before the messaging system becomes a hindrance instead of a business advantage? The means of translating business objectives and requirements for any system is usually provided in the form of SLAs. SLAs come in various forms and can be established for every facet of system function and operation. For an Exchange deployment, SLAs help management determine whether the system is delivering the services that the organization desired when investing in the system in the first place. SLAs also assist system managers in ensuring that the business services they have committed to are being delivered. SLAs come in many forms and can touch many areas of operation for an Exchange deployment. Table 2.3 shows the types of SLAs that can be established for message systems. Table 2.4 highlights some common SLAs that are used for messaging systems to ensure optimal functionality and maximum availability.
SLA Type | Description |
---|---|
Performance SLA | Established to measure how the system performs and to provide an acceptable user experience. Examples include message delivery, response time, replication periods, message rates, bytes rates, and so forth. |
Recovery SLA | Established to provide guidelines for maximum tolerable downtime. Can be set for server, mailbox, or even message recovery. Also may measure system failover times. Can also be used to drive or set recovery windows to disaster-recovery operations. |
Security SLA | Established to provide goals for ensuring system security. Can focus on intruder detection, viral and denial-of-service detection and attacks, or encryption and other secure service performance. |
Management or Administrative SLA | Established to measure administrative and management activities such as user maintenance and account service. May include service timeouts or response time guidelines. |
SLA | Description | Example |
---|---|---|
Message delivery time | Established on a site and/or organizational basis to ensure that message delivery does not exceed limits set by business requirements | 95% of messages delivered within 10 minutes (global); 99% of messages delivered within 30 minutes (global) |
Message rate | Sets a message delivery or transaction rate for the messaging system | 100 message open/second; 50 messages sent/second |
Information store recovery SLA | Maximum tolerable time that a database can be off-line | 4-hour SLA |
Individual mailbox SLA | Maximum tolerable time that a mailbox can be unavailable; also could be a recovery target SLA | 1 hour |
Gateway throughput SLA | Sets a target SLA for an SMTP or X.400 gateway in terms of messages per hour, day, and so forth | 1,000,000 messages/day; 1 GB/ hour |
Virus detection SLA | Sets a guideline for detecting and eliminating viruses | 99% viral-detection rate |
Replication SLA | Sets a guideline for ensuring replication activities such as public-folder replication, free/busy, and so forth | 24-hour replication worldwide |
Client service SLA | Target for client service availability | < 100 lost opportunities per 1,000,000 operational hours |
Server availability SLA | Target for messaging server uptime per measured period | 99.99% availability |
SLAs are the only means we have of ensuring that the messaging system is delivering business value to the organization. Without SLAs it is difficult to provide consistent functionality and service availability. Also, system managers have no criteria on which to base staffing, procedures, and training. The obvious goal should be to provide messaging services as close to 100% of the time as possible. However, uptime is not the only focus of SLAs. We can also use SLAs to provide a high degree of user experience and functionality by setting SLAs for performance, recovery, and administration. SLAs are an essential part of helping to define your operational procedures for your system. SLAs tie closely to proactive management and administration techniques and practices. We will discuss SLAs in more detail in Chapter 10.