High Availability with Unified Messaging

Multi-computer-systems and monitoring (part 2)

For a high availability concept of a unified messaging system, the message store as well as the message transport plays a decisive role. Unified messaging solution that make use of an own message store, also have to be protected separately also from the messaging server. With a real Single Store solution, at that the UM-messages for example are directly stored at the Microsoft Exchange Server, however, the high availability mechanisms for the e-mail store automatically take effect for the UM-messages as well. Independent of the storage location, the respective message, however, must be available as fast as possible in case of a breakdown, both in incoming and in outgoing direction. Proprietary concepts like e.g. the message transport via MAPI offer only limited access possibilities in case of a failure, currently present messages in queues may be excluded from the processing. When the message transport is handled e.g. via SMTP – which is common Internet technology, procedures like “Round-Robin“ (by means of setup of MX Records of the same priority in the DNS) can be implemented. When a UM-system fails, the messaging server automatically passes on the jobs at issue to the remaining, functioning UM-servers. Further established methods are the allocation of several A-Records in the DNS or load balancing systems.


It is important especially in the context with unified messaging solutions that not only the outgoing, but also the incoming traffic is made highly available by redundancy. So-called packs – this means a group of channels, e.g. realized with multiple ISDN S2M or S0 connections – are applied here. Depending on the implementation, the distribution of the incoming traffic to the different channels of the pack can be handled linearly or cyclically.


By means of a monitoring tool, the unified messaging features can be controlled constantly. In addition to that, the administrator can automatically be informed about a state of emergency, e.g. by e-mail

MTTR – the lower the better

Beside the increase of the MTBF, a MTTR as low as possible contributes to a high availability. Similar as with the MTBF, you also have to differentiate between hardware and software components of a unified messaging system here. With the hardware, so-called “Hot Plug“-components, e.g. RAID-systems with hot plug-capable hard disks are applied. A low MTTR with the software components can be achieved by monitoring systems. Unified messaging manufacturers hence face the challenge to offer particular monitoring tools, which enable the administrator to control processes, services and queues of the unified messaging solution constantly. That way, problems can be detected quickly and the MTTR can be reduced. The monitoring thereby can either be done locally on every machine or remotely from a central site. In detail, the monitoring tool fulfills the following tasks: First of all, the immediate notification of the administrator by e-mail, network message or via SMS. The ideal case would be that the administrator already tries to solve the problem before the first calls of the users reach the hotline. Secondly, the executions of defined standard measures like e.g. the restart of services. An error recovery here can happen without the intervention of the administrator. For large enterprises with a high message volume, the monitoring of unified messaging solutions is a reasonable enhancement to the redundancy concepts described above. For companies that want to implement no unified messaging system in parallel operation for reasons of economy, this is a cost-effective variant. By the automatic restart of a service as well as the immediate notification of the administrator, the downtime of the unified messaging system can be reduced and a higher availability can be achieve that way.


High availability is not only any issue for the basic business processes of a company. As unified messaging in the companies has gained in importance in the past few years, the protection from breakdowns and therewith interruptions with the transfer of messages is an essential issue. As every company has its own requirements and specifications concerning the availability of its systems, unified messaging solutions should fit in this strategy smoothly. Unified messaging solutions that can be integrated in the whole company structure including the high availability concept smoothly and that are oriented towards standards such as SMTP proved to be most reliable. Systems in parallel dynamic operation here are an effective variant, a well functioning monitoring with a consequent control of the system and the possibility to react immediately rounds off the strategy.

Copyright: serVonic GmbH
As of: September 2006

1 | 2


ImprintTerms & Conditions | © 2018 estos GmbH