Scheduling methods need to provide bandwidth, loss, and delay guarantees to users. The selected method should also provide isolation so that misbehavior by one user does not affect others. A number of scheduling methods from literature will be compared. These include: FCFS, round robin, priority queueing, priority queueing with windows, virtual clock, weighted fair queueing, self-clocked fair queueing, stop and go, and rate controlled service descipline. The complexity, advantages, and disadvantages of each of these algorithms are compared. Special issues in scheduing design for ATM networks are also discussed.

This talk covers the following topics:

• A Survey of Scheduling Methods
• Scheduling: Goals
• Goals (Cont)
• Scheduling: Methods
• Scheduling Metrics
• Scheduling Classification
• Work Conserving vs Nonconserving
• Sorted Priority vs Frame Based
• Control Distortion vs Accomodate distortion
• FCFS
• Round Robin
• Priority Queueing
• Priority Queueing with Windows
• Priority with Windows (Cont)
• VirtualClock
• VirtualClock (Cont)
• Generalized Processor Sharing
• GPS Example: Arrivals
• GPS Example: Service
• Weighted Fair Queueing
• Weighted Fair Queueing (WFQ)
• WFQ (Cont)
• GPS Example 2: Arrivals
• GPS Example 2: Service
• WFQ: Service
• Worst Case Fair Weighted Fair Queuing (WF2Q )
• WF2Q: Service
• Self-Clocked Fair Queueing (SCFQ)
• Stop and Go
• Rate Controlled Service Disciplines
• RCSD (Cont)
• Multi-class Scheduling For ATM
• Scheduling Methods: Comparison
• Summary
• References

Presentation slides in Adobe Acrobat Format: 1 slide/page