BlackBerry Enterprise Server integrates with the enterprise messaging and collaboration systems to provide mobile users with highly secure access to email, calendar, voice, instant messaging, browser, enterprise applications, and personal information management tools. It also provides advanced security features and access to administrative tools that simplify management and centralize control of the BlackBerry Enterprise Solution. With BlackBerry Enterprise Server v5.0, the key components and services, including the agent, policy service, controller, dispatcher, and router – can be configured to be highly available.
High Availability Architecture
The BlackBerry Enterprise Server high availability solution is based on a component level architecture which includes a primary server and a standby server. Either server is capable of running the mobile solution without the other. The solution is designed to support deployment options on both physical and virtual hosts (specifically VMware®), as well as a mix of both. When the BlackBerry Enterprise Server is configured for high availability, the IT administrator installs a primary BlackBerry Enterprise Server and a standby BlackBerry Enterprise Server on different computers. These BlackBerry Enterprise Server instances create a BlackBerry Enterprise Server pair. Both BlackBerry Enterprise Server instances use the same SRP credentials and BlackBerry Configuration Database. An additional BlackBerry Enterprise Server license is not required for the standby instance. Since both the primary and the standby server share the same SRP, lockouts are avoided.
With the component level architecture, health metrics are continually monitored by the BlackBerry Enterprise Server. BlackBerry administrators can set failover thresholds, which when exceeded, are designed to trigger the BlackBerry Enterprise Server to automatically switch over to the standby server. For example, if the primary server loses its connection to the mail server, automatic failover would occur to the standby server, helping to eliminate the delay of switching over manually. The administrator acknowledges when an automatic failover has occurred, fixes the problem on the originating server, and then manually sets the systems back, ensuring that failover loops are avoided.
Both BlackBerry Enterprise Server instances in the BlackBerry Enterprise Server pair include, by default, the BlackBerry Attachment Service, BlackBerry Dispatcher, BlackBerry® Mobile Data System (BlackBerry MDS) Connection Service, BlackBerry Messaging Agent, BlackBerry Policy Service, BlackBerry Router, and BlackBerry Synchronization Service components. By default, if the BlackBerry administrator chooses to install the BlackBerry Collaboration Service on both instances, the BlackBerry Collaboration Service is included in the BlackBerry Enterprise Server pair. To administer the BlackBerry Enterprise Server pair, you can install the BlackBerry Administration Service with both BlackBerry Enterprise Server instances and configure high availability separately for the BlackBerry Administration Service. In a large-scale environment, any number of BlackBerry Enterprise Server pairs that use the same BlackBerry Configuration Database can be added.
How does the high availability architecture for the BlackBerry Enterprise Server compare with clustering? To achieve high availability with server clusters often requires specialized hardware. No special hardware (such as special network cards) or software is required to achieve high availability with the BlackBerry Enterprise Server v5.0. It is built into the server software at the application (software) level.
The BlackBerry administrator can also choose to manually failover the primary server for planned service interruptions.
In load balancing, each component is configured with a virtual IP address. When a request is made on this virtual IP address, a network driver intercepts it. The request is then rerouted to one of the distributed components according to rules set by the IT administrator. In this way, the BlackBerry environment can handle increases in traffic by distributing the load among multiple components.
Other distributed BlackBerry Enterprise Server components, including the BlackBerry Attachment Service, BlackBerry MDS Connection Service Pull, and BlackBerry Router – can achieve high availability through the use of shared pools. For example, the administrator can configure a list of routers the BlackBerry Dispatcher can connect to. When the primary router fails, the dispatcher chooses a secondary router to utilize.
Setting Failover Thresholds
BlackBerry Enterprise Server components calculate a health score that indicates how well the components can provide specific services. The components send their health scores to the BlackBerry Dispatcher, which combines the health scores of the components to calculate the overall health score of the BlackBerry Enterprise Server. The BlackBerry Dispatcher writes the information to the BlackBerry Configuration Database, and it provides the information to a BlackBerry Enterprise Server that requests it.
By default, the threshold levels are set so that if the primary BlackBerry Enterprise Server loses its connection to the BlackBerry infrastructure or messaging server connection, or cannot browse the Internet, it is designed to fail over to the standby automatically. For this automatic fail-over to take place, the standby server must be connected to the BlackBerry Infrastructure and messaging server.
Log files record the activity of BlackBerry Enterprise Server components. These log files are saved on the computer that hosts the BlackBerry Enterprise Server and can be used to troubleshoot issues with component performance. With the BlackBerry® Monitoring Service, BlackBerry Enterprise Server Alert Tool, or another SNMP monitoring tool, IT administrators can monitor their BlackBerry environment for automatic fail-over events and receive notification when a switch from the primary to standby server takes place.
When the primary and standby server switch roles, both servers write the time and reason in the log files for the BlackBerry Dispatcher, Controller, and Messaging Agent, while the BlackBerry Enterprise Server Alert Tool issues an SNMP alert.
Maximizing Configuration Database Availability
In the high availability configuration, both the primary and standby BlackBerry Enterprise Servers are connected to a BlackBerry Configuration Database. Although this database can reside on any server, it is recommended that for high availability, the Configuration Database is stored on a remote SQL server and not on the same machine as the primary server. In the event of a hardware failure in the primary BlackBerry Enterprise Server, the standby server may be unable to access the Configuration Database and therefore unable to promote itself to the primary server. SQL 2005 mirroring duplicates the configuration database in real-time, as long as the same version of Microsoft SQL Server is installed on both the primary database and the mirrored configuration database. The same permissions must also be configured on both databases.
SQL 2005 mirroring requires the use of a third server instance, known as the witness. This witness is an optional instance of SQL Server than enables the mirror server to recognize whether to initiate an automatic failover. Unlike the two partners, the witness does not serve the database. Supporting automatic failover is the only role of the witness. Should the principle configuration database fail, the BlackBerry Enterprise Server components are designed to automatically connect with the mirrored database.
Planned Service Interruptions
High availability is designed to protect mobile users against both planned and unplanned service interruptions. Unplanned service interruptions are typically triggered by technical problems within the system. Planned service interruptions, such as upgrades, are often necessary for maintaining peak performance and service. The high availability architecture in the BlackBerry Enterprise Server provides the ability to perform limited-downtime upgrades. While users are still being serviced by the primary server, the standby server can be upgraded to a new version of software. The administrator can then perform a manual failover to the standby server in order to upgrade the primary server to the new version of software. This flexibility results in minimal impact to the users during server software upgrades. In addition, it provides a back-out strategy where an IT administration can manually fail back to the server running the original version if required.
Conclusions and Recommendations
Built into BlackBerry Enterprise Server, the high-availability is designed to minimize downtime from both planned and unplanned service interruptions, helping to ensure that mobile users have continuous access to e-mail, instant messaging, and other critical applications, systems and data they require to do their jobs effectively. Over 28.5 million people use BlackBerry smartphones with approximately 475 carriers and channels in over 165 countries around the world. Wireless technology gives an organization the potential to realize a clear competitive advantage through improved communication, responsiveness and productivity. Make sure to choose a solution that was built for the enterprise and can provide peace of mind – one that is trusted by some of the highest security conscious organizations in the world. High availability of the BlackBerry Enterprise Server helps maximize mobile worker productivity, improve customer service, and reduce calls to the IT help desk.