Resource Provisioning and Bandwidth Brokering for IP-core Networks Chen-Nee Chuah ISRG Retreat Jan 10 -12, 2000 Problem: How to provide end-to-end Qo. S in IP-core networks in a scalable manner?
Example Scenario H 1 H 3 ISP 2 SLA ISP 2 ISP 1 ISP 2 SLA Resource Reservation ISP 2 ISP 3 H 2 • SLA: Agreements that describe the volume of traffic exchanged, bandwidth reserved and price
Research Issues • Resource Provisioning – How to estimate bandwidth usage in advance for capacity planning purposes? • Adaptive Reservations – How to adapt aggregate reservations based on traffic fluctuation? – What are the trade-offs between granularity, Qo. S and signaling complexity? • Admission Control – End-to-end? – In stages: Per ISP cloud? Per domain?
Hierarchical Clearing House Approach destination source Edge Router ISP n ISP 2 ICH ISP 2 ISP 1 ICH CH 1 CH 2 • Distributed database – reservation status, % link utilization, optimum path • Bandwidth brokering software agent – adapt reservation dynamically
Resource Reservation Strategies • • Aggregation of reservation requests Hierarchical approach De-couple notifications & reservation requests Static and Dynamic Advanced Reservations Adapt Reservations Notifications (every Du s) - Reservation status - Link utilization - Bandwidth predictor ICH CH 2 - Advance reservations - Process reservation requests CH 1 ICH ICH ERs aggregate reservation requests (Ta)
Traffic Predictors • Monitoring system at Edge Router – Online measurement of aggregate rate of incoming & outgoing traffic (regular interval: West) • Two Traffic predictors for advanced reservations – Local Gaussian predictor for static reservation • Larger time-scale (e. g. an hour) • • Compensate for the coarse granularity of the notifications – Auto-regressive predictor for dynamic reservation • Smaller time-scale (West) •
Evaluation • Overall Performance Metrics – Link utilization – % blocking/dropping • Bandwidth Estimator – How well does the predictor track the traffic fluctuation? – Choice of estimation window, % over-provisioning • Signaling between CHs – Sensitivity analysis: effect of aggregation on Qo. S and complexity • Completely de-coupled notifications • Limited notifications
Simulation Study: Network Topology • v. BNS Backbone Network Map (1999) Boston Chicago Seattle NY Denver St. Louise SF LA • Extreme cases - Dumbbell Houston DC Atlanta Orlando - Highway with merging paths
Simulation Study: Workload Modeling • Two Qo. S classes – High priority voice calls and video conferencing – Best-effort data traffic (e. g. web, telnet, ftp) • Traffic model – Voice & video conferencing calls • Poisson arrivals with lv and lc Erlangs • Exponentially distributed call duration (mean = 2. 5 min. for voice, 30 min. for video conferencing calls) • Individual source is modeled as two state-Markov chain. When “on”, a voice call requires bandwidth of 128 kbps, defined as one basic unit (BU) • Video conferencing calls occupy 4 BU – TCP connections get equal share of the non-reserved bandwidth
Скачать презентацию Resource Provisioning and Bandwidth Brokering for IP-core Networks
3e2cbd81286e0a15109d43d11e80daed.ppt