Q2. What features will need to be added to the PMIF?
Features supported by some tools, such as priority scheduling and domains, that can be solved with approximate solution algorithms will also need to be added.
Q3. What is the way to handle software model data interchange?
The best way to handle software model data interchange is to use the SPE meta-model [WILL95] with a transfer format derived from it, rather than adding software model extensions to the PMIF.6.0 StatusThe original proposal was the result of a research project.
Q4. What are the extensions needed for the PMIF?
1997 by the authorsSeveral extensions are needed: the graphical representation of the model, the experimenter, the studies, the results, and additional model features such as priorities, domains, passive resources, scheduling disciplines, and so on.
Q5. What is the purpose of the PMIF?
Terminology used in the PMIF is based on terms that are meaningful to developers of performance modeling tools rather than terms familiar to tool users.
Q6. What are the resources required for the servers in the model?
The software model solution yields the following resource requirements for the servers in the model:• Withdrawal: (630, 11, 8)• Get Balance: (250, 6, 3)for the CPU, ATM, and DEV1 respectively.
Q7. What is the way to represent the QNM picture?
The authors could adopt the CDIF© 1997 by the authorsparadigm, however, it appears that the QNM picture could be handled with additional attributes attached to nodes to define “a virtual grid” location, and a picture-element type.
Q8. What was the need for a mechanism for exchanging information about VSLI designs among chip?
They needed a mechanism for exchanging information about VSLI designs among chip foundries and research and production tools that conducted a variety of design analyses and error checking functions.
Q9. What is the value of the arc entity in the PMIF?
Each of the entities and its attributes is represented as follows:(EntityName EntityID (Attribute1 Attribute1Value) … (AttributeN AttributeNValue) )For example, an arc entity in a queueing model may be defined as: (Arc QNM001.1(FromNode #d4) (ToNode #d3))Arc is the EntityName and the EntityID is the Meta-entityID followed by an InstanceNumber, i.e., QNM001.1.
Q10. What is the way to handle the experimenter specifications?
After reviewing the requirements, it appears that it is possible to handle the experimenter specifications as a meta-model extension, and to tie the model results to the experimenter rather than the QNM meta-model.
Q11. What are the three basic approaches to specifying service requests for servers?
There are three basic approaches to specifying service requests for servers in the model:- Time: specifies the service time per visit to a server- Demand: specifies the total service time for all visits to a server- Computed: calculates either time or demand from other data specified in the model.© 1997 by the authors•