Stateless core : a scalable approach for quality of service in the Internet : winning thesis of the 2001 ACM Doctoral Dissertation Competition /
Saved in:
| Author / Creator: | Stoica, Ion, 1965- |
|---|---|
| Imprint: | Berlin ; New York : Springer, 2004. |
| Description: | xvi, 219 p. : ill. ; 24 cm. |
| Language: | English |
| Series: | Lecture notes in computer science, 0302-9743 ; 2979 ACM distinguished theses ACM distinguished dissertations. |
| Subject: | |
| Format: | E-Resource Print Book |
| URL for this record: | http://pi.lib.uchicago.edu/1001/cat/bib/5169996 |
Table of Contents:
- 1. Introduction
- 1.1. Main Contribution
- 1.2. Other Contributions
- 1.3. Evaluation
- 1.4. Discussion
- 1.4.1. Why Per Flow Processing?
- 1.4.2. Scalability Concerns with Stateful Network Architectures
- 1.5. Organization
- 2. Background
- 2.1. Circuit Switching Vs. Packet Switching
- 2.2. IP Network Model
- 2.2.1. Router Architecture
- 2.2.2. Data Path
- 2.2.3. Control Path
- 2.2.4. Discussion
- 2.3. Network Service Taxonomy
- 2.3.1. Best Effort Service
- 2.3.2. Flow Protection: Network Support for Congestion Control
- 2.3.3. Integrated Services
- 2.3.4. Differentiated Services
- 2.4. Summary
- 3. Overview
- 3.1. Solution Overview
- 3.1.1. The Stateless Core (SCORE) Network Architecture
- 3.1.2. The "State-Elimination" Approach
- 3.1.3. The Dynamic Packet State (DPS) Technique
- 3.2. Prototype Implementation
- 3.2.1. An Example
- 3.3. Comparison to Intserv and Diffserv
- 3.3.1. Intserv
- 3.3.2. Diffserv
- 3.4. Summary
- 4. Providing Flow Protection in SCORE
- 4.1. Background
- 4.2. Solution Outline
- 4.3. Core-Stateless Fair Queueing (CSFQ)
- 4.3.1. Fluid Model Algorithm
- 4.3.2. Packet Algorithm
- 4.3.3. Weighted CSFQ
- 4.3.4. Performance Bounds
- 4.3.5. Implementation Complexity
- 4.3.6. Architectural Considerations
- 4.3.7. Miscellaneous Details
- 4.4. Simulation Results
- 4.4.1. A Single Congested Link
- 4.4.2. Multiple Congested Links
- 4.4.3. Coexistence of Different Adaptation Schemes
- 4.4.4. Different Traffic Models
- 4.4.5. Large Latency
- 4.4.6. Packet Relabeling
- 4.4.7. Discussion of Simulation Results
- 4.5. Related Work
- 4.6. Summary
- 5. Providing Guaranteed Services in SCORE
- 5.1. Background
- 5.2. Solution Outline
- 5.3. Data Plane: Scheduling without Per Flow State
- 5.3.1. Jitter Virtual Clock (Jitter-VC)
- 5.3.2. Core-Jitter-VC (CJVC)
- 5.3.3. Data Path Complexity
- 5.4. Control Plane: Admission Control with no Per Flow State
- 5.4.1. Ingress-to-Egress Admission Control
- 5.4.2. Per-Hop Admission Control
- 5.4.3. Aggregate Reservation Estimation Algorithm
- 5.5. Experimental Results
- 5.5.1. Processing Overhead
- 5.6. Related Work
- 5.7. Summary
- 6. Providing Relative Service Differentiation in SCORE
- 6.1. Background
- 6.2. Solution Outline
- 6.3. LIRA: Service Differentiation Based on Resource Right Tokens
- 6.3.1. Link Cost Computation
- 6.3.2. Path Cost Computation and Distribution
- 6.3.3. Multipath Routing and Load Balancing
- 6.3.4. Route Pinning
- 6.3.5. Path Selection
- 6.3.6. Scalability
- 6.4. Simulation Results
- 6.4.1. Experiment Design
- 6.4.2. Experiment 1: Local Fairness and Service Differentiation
- 6.4.3. Experiment 2: User Fairness and Load Balancing
- 6.4.4. Experiment 3: Load Distribution and Load Balancing
- 6.4.5. Experiment 4: Large Scale Example
- 6.4.6. Summary of Simulation Results
- 6.5. Discussion
- 6.6. Related Work
- 6.7. Summary
- 7. Making SCORE More Robust and Scalable
- 7.1. Failure Model
- 7.1.1. Example
- 7.2. The "Verify-and-Protect" Approach
- 7.2.1. Node Identification
- 7.2.2. Protection
- 7.2.3. Recovery
- 7.3. Flow Verification
- 7.3.1. Bufferless Packet System
- 7.3.2. Flow Identification Test
- 7.3.3. Setting Threshold H u
- 7.3.4. Increasing Flow Identification Test's Robustness and Responsiveness
- 7.4. Identifying Misbehaving Nodes
- 7.4.1. General Properties
- 7.5. Simulation Results
- 7.5.1. Calibration
- 7.5.2. Protection and Recovery
- 7.6. Summary
- 8. Prototype Implementation Description
- 8.1. Prototype Implementation
- 8.1.1. Updating State in IP Header
- 8.1.2. Data Path
- 8.1.3. Control Path
- 8.2. Carrying State in Data Packets
- 8.2.1. Carrying State in IP Header
- 8.2.2. Efficient State Encoding
- 8.2.3. State Encoding for Guaranteed Service
- 8.2.4. State Encoding for LIRA
- 8.2.5. State Encoding for CSFQ
- 8.2.6. State Encoding Formats for Future Use
- 8.3. System Monitoring
- 8.4. System Configuration
- 8.4.1. Router Configuration
- 8.4.2. Flow Reservation
- 8.4.3. Monitoring
- 8.5. Summary
- 9. Conclusions and Future Work
- 9.1. Contributions
- 9.2. Limitations
- 9.3. Future Work
- 9.3.1. Decoupling Bandwidth and Delay Allocations
- 9.3.2. Excess Bandwidth Allocation
- 9.3.3. Link Sharing
- 9.3.4. Multicast
- 9.3.5. Verifiable End-to-End Protocols
- 9.3.6. Incremental Deployability
- 9.3.7. General Framework
- 9.4. Final Remarks
- A. Performance Bounds for CSFQ
- B. Performance Bounds for Guaranteed Services
- B.1. Network Utilization of Premium Service in Diffserv Networks
- B.2. Proof of Theorem 2
- B.3. Proof of Theorem 3
- B.3.1. Identical Flow Rates
- B.3.2. Arbitrary Flow Rates
- B.4. Proof of Theorem 4
- References
