Question

1.Briefly describe Polling based and Event based management, select the one that is better to use in practice, and provide a supporting argument for your choice. Provide an analogy of your choice.

Answer:

2.Define SNMP, briefly discuss the single most important difference between SNMPv2 and SNMPv3, and explain how this difference is achieved in the protocol that supports it.

Answer

3. What are the different management philosophies in the context of management organization. Briefly discuss each.

Answer:

4.List the main factors that best discuss the complexity that arises when quantifying management integration. Select the one factor that deals with the variety of management functions and elaborate more on this factor.

Answer:

5. List and briefly discuss the criteria that must be met about a given parameter in SLA.

Answer:

6.List and briefly discuss the classical metrics needed to assess reliability and availability of networks and services. Provide an example of the metric used to determine the time interval needed to service an equipment.

Answer:

Answer 1

1)

Polling or events are basically used to check for existence of new data or updation of existing data.

Polling:

As the name says, we constantly need to poll (check) the target for the desired checks. The interval (time frequency) to poll totally depends on how much ever we want to have. Lesser the time (higher frequency), harder the cpu has to work.

Events:

Events basically achieve the same thing as what polling achieves but in a different (efficient) way. As the name says, we depend on an event to check the target. We don't need to constantly check for the target. Instead, we depend on a event from the target to notify us for the change. Based on this event, we connect to the target and perform the necessary operation. Unlike, polling, in this case, we need not waste CPU cycles and can instead wait in sleep mode till an event is fired.

Example:

Consider the example of file systems in the context of Java programming. Lets create a usecase. Whenever a new file gets created in a directory, we need to read it and do some processing.

In case of polling, we need to constantly check the directory if there's any new file and process it.

In case of event, Java programming language has a separate API to deal with filesystem events. We register the directory for new events and write a callbacks to handle it.

By now, you might have understood the obvious benefit of events. They're efficient with almost no cpu cycle wastage; whereas in polling, all the cpu cycles are waste except the one which identifies a new file. So events are better than polling in terms of efficiency. However, not all systems support events. The bottom line is to go with events whenever they're available and to fallback to polling if events are not available.