For each set of application performance requirements shown in Figure 3.33, classify the network as single-tier or multi-tier performance. Please explain your choice for each.
PERFORMANCE REQUIREMENTS
APPLICATION SET
Capacity
Reliability
Delay
Application Set 1
Application 1 150 Kb/s
99.95%
100 ms
Application 2 100 Kb/s
99.95%
100 ms
Application 3 100 Kb/s
99.95%
100 ms
Application 4
120 Kb/s
99.95%
100 ms
Application Set 2
Application 1 75 Kb/s
99.999%
40 ms
Application 2 150 Kb/s
99.95%
100 ms
Application 3 250 Kb/s
99.999%
40 ms
Application 4 200 Kb/s
99.95%
100 ms
Application Set 3
Application 1 1.1 Mb/s
99.95%
40 ms
Application 2 800 Kb/s
99.95%
100 ms
Application 3 950 Kb/s
99.95%
100 ms
Application 4 120 Kb/s
99.95%
20 ms
Application 5
120 Kb/s
99.999%
50 ms
Figure 3.33 Application Performance Requirements for Exercise 3.2 …show more content…
Application Set 1:
This Application set could be classified under a single tier network.
As seen from the table, the reliability and delay is the same for all four applications and the capacity of each of them does not differ from each other a lot with the values ranging from 100 Kb/s to 150 Kb/s. Hence all applications could use a single tier network.
Application Set 2:
This application set could be classified under a multi network. The applications 1 and 3 have one set of reliability and delay requirements while applications 2 and 4 have another set. The capacity of applications 1 and 3 vary quite a lot hence it has to be configured under different tiers. Application 2 and can be configured in one.
Application Set 3
The application Set 3 requires a multi-tier network as the requirements vary a lot from one application to another. Application 1 could be classified in one tier and applications 2 and 3 in one as their requirement is quite similar. Applications 4 and application 5 can be a part of 2 more tiers.
Problem
4.6
You are developing a network for a company’s on-line transaction processing (OLTP) application (e.g. a retail sales network). Their current system is a mainframe that has several terminals connected to it, either directly or through a terminal server, as in Figure 1 below. They are moving to a hierarchical client server network, where there will be multiple regional database servers, each acting in a client-server fashion and updating each other’s regions via a database manager, as in Figure 2.
Show the probable data sources and sinks for both environments.
Figure 1 - Mainframe Environment for OLTP Application
Figure 2 - Hierarchical Client Server Environment for OLTP Application
a) How does migrating from the mainframe environment to the hierarchical client-server environment modify the traffic flows in the network?
In the mainframe environment, the dataflow is quite predictable and straightforward. The user devices will send requests to the mainframes or the terminal server. The entire request from the terminal server is processed and data is sent to the mainframe server and the response will flow out of the mainframe. The mainframe is connected to the database and the processed data will be stored in the database.
In the second case, the dataflow is highly unpredictable. Data may be needed to be pulled out of any of the database servers. Hence the manager will have to perform several routing if the user device is not able to access information from the database it is connected to.
b) In what ways does the network environment improve the traffic flows?
In the mainframe network, there is a lot of pressure on the mainframe to manage all the data and traffic flow. The traffic is managed better in the Hierarchical Client Server Environment. The user devices are connected directly to databases. A lot of traffic would be reduced on account of this. Also other traffic between the databases can be balanced better in this network and hence traffic management is a lot better.
c) What are some of the potential trade-offs between the two environments, for example, in security, management, and performance?
The Security in mainframe environment is better as the security is to be given to just one server. In the distributed server, the security has to be distributed across systems.
Management of data is simple in mainframes as the mainframe does all the work but it is not very effective. Management is trickier in distributed systems but much more effective if the architecture is planned well. More data can be managed by the distributed systems than on mainframes.
Performance of a distributed system is much higher compared to the mainframes as the data and traffic flow is spread across the entire network as compared to just the mainframe in the mainframe environment.