High Availability with Unified Messaging

Multi-computer-systems and monitoring

by Iris Walter | serVonic GmbH | Public Relations


The operation of unified messaging solutions is already part of the daily working routine in many companies, because with unified messaging, a consistent communication platform for e-mail, fax, SMS and voice mail is available. This does not only save money for the company, but also precious work time for employee. An important factor in the company-own communication thereby is the availability of the whole system. There is hardly a company that can afford to be cut off from the outside world on account of the failure of single components. It surely varies how urgent, important and indispensable the messages – incoming and outgoing – of a company are. As a result, every company has to develop and implement a suitable strategy in order to guarantee at least a sufficient availability of the communication systems in the company. Where high availability is concerned, a unified messaging solution should integrate in this strategy properly, because if the unified messaging solution fails, the employees are cut off from essential communication channels abruptly. And an uninterruptible course of business is a decisive competition factor for companies.

Developing strategies and implementing them

High availability for a complete system like the company communication with all the hardware- and software-components involved can be achieved in two ways: One way is to set up the complete system redundantly: When the system fails in case of a breakdown or because of service works, the remaining systems take over the necessary processes. The other way is to plan each of the single components itself highly qualitative, so that a failure is as unlikely as possible. This means that components are applied that for example already have passed long-term tests successfully or software- and hardware-components from reliable manufacturers, that are well-established in the market, are deployed. Usually, a mix of both the variants – redundancy and highly reliable single components – is applied in order to achieve a high availability of the complete system.

High availability – a definition

Basically, a system is called highly available when it can fulfill the tasks that it is meant for. The availability of a system is stated in percentage quotation, that is calculated from the ratio of “downtime“ and “uptime“ of the system within a year:

Availability = Uptime / (Downtime + Uptime)

Whereas some IT-specialists consider a system with 99 percent – this means a downtime of 3.6 days – as highly available, others fix a threshold value of 99.99 percent; this means a downtime of 52 minutes. Apart from the “Mean Time Between Maintenance“ (the planned downtime like e.g. maintenance work), the downtime mainly depends on the “Mean Time Between Failure“ and the “Mean Time To Repair“. Mean Time Between Failure – or MTBF – means the average period of time between two failures. Mean Time To Repair – or MTTR – describes the average period for the recovery of the system after a failure. In the following, different technical concepts shall be presented in order to have a “positive” influence with the implementation of unified messaging systems.

When a unified messaging server fails, the messaging system automatically passes on the jobs at issue to the remaining, functioning UM-servers

Increase of the MTBF

With the implementation of a highly available unified messaging system, a comprehensive view at the all the components applied – hardware as well as software – is essential. As the complete availability is a result of the single availabilities, every single component has to meet the claims concerning the availability. Whereas single hardware components like fans, hard disks or network interface cards can be designed redundantly also with a single-server implementation, this cannot be realized with the software, consisting of operating system and services, on one single PC. A common and proved concept in order to create the necessary redundancy in this case and that way manage to achieve a high MTBF is the application of computer clusters. For this, the unified messaging solution is installed on multiple computers identical in construction, this means with identical capacities in channels and services, that are networked with each other. From the outside, the networked computers appear as one system. When one unified messaging server fails, the message load so can be handled via the other servers. Depending on the architecture, you can make a distinction between standby-systems and systems in parallel operation. With a standby-system, the technology of the Active-/Passive-Clustering is applied. During the normal operation, only one server processes the complete volume of messages, the second server is available passively for the failover. The period of time the unified messaging systems are down is reduced to the time to switch over here. Services and features of the failed server are available to the user again in shortest time. Depending on the point of view – technological or user-oriented – the short-term fail of the unified messaging system is integrated in the calculation of the MTBF. With systems in parallel operation, a time to switch over is completely out of question. In addition to that, the available resources here are used optimally also in normal operation. Unified messaging systems in parallel operation are based on the technology of the Active-/Active-Clustering. All the unified messaging systems installed run actively here and process the message volume in parallel. When a system fails, the remaining systems take over the processing of the jobs in the queue. The responsibility assignment between the systems is handled according to the “Load-Balancing“-principle. [more] 

 


1 |

 

Information

Find more information on high availability and monitoring in our datasheets.

High availability


Monitoring

Whitepaper

Contact

Iris Walter
+49 8142 4799-29

Please let us know if you need more information on other whitepapers.

ImprintTerms & Conditions | © 2018 estos GmbH