This space can be utilized for clearing doubts on the concepts of Advanced Operating System PECS 3409 of BPUT Syllabus. Questions will be added to the topic body and answers will be discussed in the comments.
- What is consistency maintenance in distributed system and what are its prime categories?
- Discuss the technique by which RPCs can be replaced in threaded environment for increasing performance?
- What is Omega Network?
- Draw 16 input Omega network
16 input Omega Network
See this document for other Answers :
Q 1 Discussion
Consistency as a problem arises due to (a) off-line storage of data – in case of caching or (b) due to the distributed nature of the system.
Consistency maintenance is about enforcing consistency rules in the aforesaid scenario. Consistency can be obtained by adopting two methods (a) Polling (b) Notification.
Polling requires active request by a distributed node (Si) to (Sj) requesting updated data. Notification requires the changer of the data to publish its changes by contacting the other nodes using / sharing the same data.
However depending upon the nature of the system, the consistency needs may not be the same. (Refer Consistency models)
Q 2 Discussion
RPCs are intended for deployment in a remote deployment scenario but they are heavily used in single system deployments too because of their streamlined design and ease of deployment. In these cases RPC calls target the same system. This example is a perfect candidate for replacement. In a threaded environment, the RPCs can be replaced by the following steps :-
1. Removal of the stub procedures
2. Removal of advertised services
3. Replacement of stub procedures with actual function calls.
There are other alternatives available also :-
1. Creation of Daemon processes on the same host which can spawn multi-threaded procedures on demand.
2. Creation of multi-threaded Shared libraries
Q 3 Discussion
The crossbar interconnection network is scalable in terms of performance but unscalable in terms of cost. Alternatively the shared bus network is scalable in terms of cost but unscalable in terms of performance. An Intermediate class of networks named multistage interconnection networks lies between the two extreme varieties. It is more scalable than the bus for performance and more scalable than the crossbar in terms of cost.
An Omega network is a multistage interconnection network configuration often used in parallel computing architectures. It is an indirect topology that relies on the perfect shuffle interconnection algorithm. It consists of log p stages where p is the number of processing nodes and also the number of memory banks. The omega network has p/2Xlogp switching nodes and the cost of such a network grows as theta(p log p). The cost is less than theta(p square) the cost of a complete crossbar network.