Problems with Traditional Systems
From past to today, traditional telephony system architectures are:
- Centralized systems with single CPU or server: Systems with a well-defined centralized control process
- Centralized systems with redundant CPU or server: Providing some extent protection against total system failure
- Distributed systems : A collection of subsystems that appears to the application as a single coherent system. Still a CPU or server exists in order to provide the integrity
- Distributed systems with the redundant CPU or server that might be geographically distributed : Providing more protection against total system failure
Although the above mentioned architectures achieved the main objective of delivering defined telephony functions with some cost, employing all these might still create a significant headache to some developers and system owners. There are inherent problems with these traditional architectural solutions like:
- Increasing the solution capacity might be costly due to the need for more processing power and memory
- Addition of redundant CPU or server required in order to decrease the risk of total system failure. That is an additional cost
- Architectural complexities due to many hardware and software components
- No possibility to erase the risk for total system failure since there are always critical components (critical component here states any component which may cause total system failure)
Solution: Telesis CAU
To deal with the mentioned issues, Telesis A.S. has a new approach that is CAU – Cluster of Autonomous Units. In this approach, autonomously acting individual units perform their local task but also participate on a global task. Individual unit has minimal hardware, and is as simple as possible. No central processing unit or server exists at all.
The core of the Telesis CAU and the essence of its technological sophistication is clustering autonomous units, which includes:
- No central control at all. No CPU or servers. Each unit in the cluster has its own local processing power
- No critical components, which may lead to total system failure. Faults are local, do not propagate. Fast, autonomous reaction to problems
- The system database saved in each autonomous unit. Almost no possibility of loosing the database
- Increase of system processing power with the capacity expansion. No performance degradation if more units are added to the system
- Geographic distribution of autonomous units. Capability to protect the system and its components against unwanted or unexpected environmental influences
- System performs as requested regardless of the number of units
Standalone Autonomous Unit Models:
Autonomous Unit Models in Board (Card) Form:
