Resource Negotiation, Pricing and QoS for Adaptive Multimedia Applications Xin Wang Work with Henning Schulzrinne Internet Real -Time Laboratory Columbia University http://www.cs.columbia.edu/~xinwang/RNAP.html

Scope of Metropolitan IP Network Service Level Agreements (SLA) are negotiated based on Application Specific Needs bandwidth, loss, delay, jitter, availability, price Application SLA Large number of new applications are appearing in the Internet. This includes the real-time audio, video, and mission-critical financial data. This provides ISP more business opportunity, and also challenge. The value-added services normally require certain service expectations. Since different applications have different requirements in bandwidth and quality, network resource provision is challenging. Proliferation of new IP services and applications with different bandwidth and quality of service requirements presents significant opportunities and challenges for service providers for IP networks provisioning. 4/23/2019

The needs from Next Generation Service Providers Increase revenue by offering innovative IP services: Delivering high-margin, differentiated services VoIP, VPN, Applications Hosting etc Gain competitive advantage by deploying new services more quickly, placing a premium on time to market and time to scale Reduce cost and operation complexity Evolve from static network management to dynamic service provisioning Reduce costs by automating network and service management Service provider needs to provide different premium services, without big cost. The network management and resource provisioning should be automatic and fast. It is critical for ISPs to offer premium IP services cost effectively to offset the declining revenue from the traditional connectivity services (raw bandwidth). 4/23/2019

Internet Structure End User LAN POP NAP Backbone Provider Regional Provider Private Network Backbone Provider NAP LAN End User Backbone Provider POP Private Peering Before deciding on the services for the internet, let’s look at the current Internet situation. Network is generally divided into different management domains, with direct peering or connect through Network Access Point (NAP). The Network Access Point (NAP) allows Internet Service Providers (ISPs) to interconnect and exchange information among themselves. The exchanging of Internet traffic is generally referred to as "peering". 4/23/2019

NORDUnet Traffic Analysis Before further describes the service policy, we show some traffic statistics of NORDUnet NORDUnet interconnects the Nordic national networks for research and education and connects these networks to the rest of the world. The current physical connections are shown on the connectivity map. NORDUnet provides its services by a combination of leased lines and Internet services provided by other international operators. 4/23/2019

NORDUnet Traffic Analysis Results: All access links (interconnect ISP’s or connect private networks to ISP’s), including trans-Atlantic links, can get congested. Average utilization is low: 20-30% Peak utilization can be high: up to 100 Congestion Ratio (peak/average): as high as 5. Peak duration can be very long: Chicago NAP congested once in 8/00, lasted 7 hours. TeleGlobe links congested every workday in 8/00 and 9/00 Reasons: Frequent re-configuration and upgrading;Load balancing to protect own users. 4/23/2019

Solutions ? Simply over-provisioning? Quality of Service (QoS)? Adding enough bandwidth for all applications Will over-provisioning be sufficient to avoid congestion? How much bandwidth is enough to meet diverse user requirements? No intrinsic upper limit on bandwidth use How much it costs to add capacity? Quality of Service (QoS)? Conditioning the network to provide some predictability to an application How efficient a QoS mechanism manages the bandwidth? How much complexity? How much a user needs to pay for QoS? Application adaptation? Adapting the source rate based on network conditions to avoid congestion Why would an application adapt? To leave room for others? Demand: Availability of more bandwidth will create its own demand through increasing utilization of bandwidth intensive applications” .real-time audio//video, 3D imaging, virtual reality, etc. Supply: Cost of transportation using fiber optics is declining drastically. However, network management cost: switches and routers, state of the art POP, data centers, etc, will cost money. QoS: Protect the valuable applications through QoS. But will QoS add big complexity? Providing different servers needs different pricing, otherwise, everyone will ask the best service and end up not services. Applications do not have the motivation to adapt. 4/23/2019

Bandwidth Pricing Reality: leased bandwidth price has not been dropping consistently and dramatically. Facts: 300 mile T1 price (rent): 1987: $10,000/month 1992: $4,000/month 1998: $6,000/month (thanks to high Internet demand) 100-mile cabling cost in 1998: $65,000 Purchase a higher-speed connection if the option to manage the existing bandwidth to support the applications is available? What company can disregard high recurring monthly costs? 4/23/2019

Bandwidth Pricing (cont.) Links connecting ISP’s are very expensive Facts: International Frame Relay with 256-kbps: thousands dollars a month. Transit DS-3 link: $50,000/month between carriers. Transit OC-3 link: $150,000/month between carriers. Chicago NAP: $3,900/month/DS-3, $4,700/month/OC-3. T1 - 1.544 megabits per second (24 DS0 lines) T3 - 43.232 megabits per second (28 T1s) OC3 - 155 megabits per second (100 T1s) OC12 - 622 megabits per second (4 OC3s) OC48 - 2.5 gigabits per seconds (4 OC12s) OC192 - 9.6 gigabits per second (4 OC48s) The price for a Chicago NAP connection is distance sensitive and based on the location where the ISP's network meets Ameritech's. ATM pricing also varies with contract length with price deductions for longer term contracts. NAP connection prices start at $3,900 per month for a DS3 and $4,700 per month for an OC3. Duration of 12, 36 or 60 month terms are available. Bandwidth may be cheap, but not free. We need a service model and pricing mechanism that would allocate bandwidth more efficiently. 4/23/2019

Solutions ? Simply over-provisioning? Quality of Service (QoS)? Adding enough bandwidth for all applications Will over-provisioning be sufficient to avoid congestion? How much bandwidth is enough to meet diverse user requirements? No intrinsic upper limit on bandwidth use How much it costs to add capacity? Quality of Service (QoS)? Conditioning the network to provide some predictability to an application How efficient a QoS mechanism manages the bandwidth? How much complexity it will add? How much a user needs to pay for QoS? Application adaptation? Adapting the source rate based on network conditions to avoid congestion Why would an application adapt? To leave room for others? Demand: Availability of more bandwidth will create its own demand through increasing utilization of bandwidth intensive applications” .real-time audio//video, 3D imaging, virtual reality, etc. Supply: Cost of transportation using fiber optics is declining drastically. However, network management cost: switches and routers, state of the art POP, data centers, etc, will cost money. QoS: Protect the valuable applications through QoS. But will QoS add big complexity? Providing different servers needs different pricing, otherwise, everyone will ask the best service and end up not services. Applications do not have the motivation to adapt. 4/23/2019

What we add to these solutions? Developed a service negotiation framework QoS support + user adaptation, allows resource commitment for short intervals. Proposed a Resource Negotiation And Pricing protocol RNAP: Enables user and network (or two network domains) to dynamically negotiate services Can be embedded in other protocols, e.g., RSVP, or implemented independently Proposed a pricing model Enable differentiated charging to suppor multiple levels of services Considers long-term user demand Allows for congestion pricing to motivate user adaptation 4/23/2019

Goals Achieve Network Efficiency Achieve Economic Efficiency Maintain a target level of service while minimizing the resources needed to provide this service Improves network availability and QoS for the users Allow a service class to meet its performance assurances under large or bursty offered loads, without high network management overhead, or over-provision cost Achieve Economic Efficiency Price services levels to reflect the QoS cost Under resource contention, allocate resources to the users that value them most Maximize the user benefit, maximize network revenue 4/23/2019

Protocol Architectures: Centralized Host Resource Negotiator RNAP Messages Network Resource Negotiator NRN NRN NRN HRN HRN Access Domain - A We consider two alternative architectures for implementing RNAP in the network, a centralized architecture (RNAP-C) and a distributed architecture (RNAP-D) In RNAP-C, user negotiates through a HRN, each network domain has a NRN. When a user wants to to apply for resources from the network, it first sends a request to the NRN of its access domain. This request is then propagated to next-domain NRN, and so on. In general, each NRN is in charge of admission control, monitoring network statistics price quotation and charging for its domain. Edge Router Access Domain - B Internal Router Intra-domain messages Transit Domain RNAP-C 4/23/2019

Protocol Architectures: Distributed RNAP Messages HRN LRN LRN LRN LRN LRN LRN LRN LRN LRN HRN LRN LRN Access Domain - A LRN LRN Edge Router Access Domain - B In RNAP-D, Local Resource Negotiators (LRN) were implemented on each router, for admission control, monitoring network statistics, forming price for each service class. At network edge, NRNs dynamically configure traffic conditioners, based on on-going user requests. Internal Router Transit Domain RNAP-D 4/23/2019

RNAP Messages Periodic re-negotiation Query: Inquires about available services, prices Query Quotation Quotation: Specifies service availability, accumulates service statistics, prices Reserve Commit Reserve: Requests/Updates service(s), resources Periodic re-negotiation Quotation Commit: Admits the service request at a specific price or denies it. Reserve Commit Close: Tears down negotiation session Close Release: Releases the resources Release 4/23/2019

Message Aggregation (RNAP-D) Turn off router alert Turn on router alert Sink-tree-based aggregation 4/23/2019

Message Aggregation (RNAP-C) Sink-tree-based aggregation 4/23/2019

RNAP Message Aggregation Summary Aggregation when senders share the same destination network Messages merged by source or intermediate domains Messages de-aggregated at destination border routers (RNAP-D), or NRNs (RNAP-C) Original messages sent directly to destination/source domains without interception by intermediate RNAP agents; aggregate message reserves and collects price at intermediate nodes/domains Overhead Reduction Processing overhead, storage of states 4/23/2019

Block Negotiation Aggregated resources are added/removed in large blocks to minimize negotiation overhead and reduce network dynamics Bandwidth time 4/23/2019

Pricing in Current Internet Access-rate-dependent flat charge (AC) Simple, predictable Difficult to compromise between access speed and cost No incentive for users to limit usage congestion Usage-based charge Volume-dependent charge (V) Time-base charge (T) work better for uniform per-time unit resource demands, e.g., phone Access charge + Usage-based charge Per-hour charge after certain period of use, or per-unit charge after some amount of traffic transmitted. Flat charge for basic service, usage charge for extra bandwidth or premium services Flat fee: Predictable fee for both users and providers. Avoid potentially considerable administrative costs of tracking, allocating and billing for usage 4/23/2019

Two Volume-based Pricing Strategies Fixed-Price (FP): fixed unit volume price FP-FL: per-byte charge are same for all services FP-PR: service class dependent FP-T: time-of-day dependent FP-PR-T: FP-PR + FP-T During congestion: higher blocking rate OR higher dropping rate and delay Congestion-Price-based Adaptation (CPA): FP + congestion-sensitive price component CP-FL, CP-PR, CP-T, CP-PR-T Congestion price calculation: Tatonnement process or auction During congestion: users maintain service by paying more OR reduce sending rate OR lower service class In period of resource scarcity, quality sensitive applications can maintain their resource levels by paying more \Quality insensitive applications will reduce their sending rate or change to a lower service class 4/23/2019

Important Time Scales Technical levels of interactions Monetary levels of interactions atomic short-term medium-term long-term Retransmission Error Handling Flow Control Resource Reservation Capacity Planning Scheduling Feedback Policing Routing time Congestion Pricing Time-of-day Flat Rates Pricing 4/23/2019

Pricing Strategies Holding price and charge: cost by depriving others’ opportunity to be admitted even if no data is sent phj =  j (pu j - pu j-1) , chij (n) = ph j r ij (n) j Usage price and charge: optimize the provider’s profit max [Σl x j (pu1 , pu2 , …, puJ ) puj - f(C)], s.t. r (x (pu2 , pu2 , …, puJ ))  R cuij (n) = pu j v ij (n) Congestion price and charge: drive demand to supply Tatonnement process (CPA-TAT): pc j (n) = min [{pcj (n-1) +  j (Dj, Sj) x (Dj-Sj)/Sj,0 }+, pmaxj ] ccij (n) = pc j v ij (n) Second price auction(CPA-AUC): charge as the highest biding price of the requests being rejected. 4/23/2019

Usage Price for Differentiated Service Usage price based on cost of class bandwidth: lower target load (higher QoS) -> higher per-unit bandwidth price Parameters: pbasic basic rate for fully used bandwidth  j : expected load ratio of class j xij : effective bandwidth consumption of application i Aj : constant elasticity demand parameter Price for class j: puj = pbasic /  j Demand of class j: xj ( puj ) = Aj / puj Effective bandwidth consumption: xe j ( puj ) = Aj / ( puj  j ) Network maximizes profit: max [Σl (Aj / pu j ) pu j - f (C)], puj = pbasic /  j , s. t. Σl Aj / ( pu j  j )  C Hence: pbasic = Σl Aj / C , puj = Σl Aj /(C j) 4/23/2019

Rate Adaptation of Multimedia System Enable multimedia applications to gain optimal perceptual value based on the network conditions and user profile. A Host Resource Negotiator (HRN) negotiates services with network on behalf of a multimedia system. Utility function: users’ preference or willingness to pay Cost U1 U2 Utility/cost/budget U3 Budget Bandwidth 4/23/2019

Utility User adaptation under CPA-TAT User adaptation under CPA-AUC Optimally distributes network-allocated among applications in a multimedia system, optimize the system performance U = Σi Ui (xi (Tspec, Rspec)] max [Σl Ui (xi ) - Ci (xi) ], s. t. Σl Ci (xi)  b , xmini  xi  xmaxi Determine optimal Tspec and Rspec User adaptation under CPA-AUC Literature model: not consider short-term resource reservation; not address user response to price change; assume unique service request; service uncertainty, signaling burst, setup delay M-bid auction model: express bid for a number of bandwidth; network select one that has the lowest bidding price 4/23/2019

Example Utility Function User defines utility at discrete bandwidth, QoS levels Utility is a function of bandwidth at fixed QoS An example utility function: U (x) = U0 +  log (x / xm) U0 : perceived (opportunity) value at minimum bandwidth  : sensitivity of the utility to bandwidth Function of both bandwidth and QoS U (x) = U0 +  log (x / xm) - kd d - kl l , for x  xm kd : sensitivity to delay kl : sensitivity to loss Optimization: max [Σl U0i + i log (xi / xmi ) - kdi d - kl i l - pi xi ], s. t. Σl pi xi  b , x  xm , d  D, l  L Without budget constraint: x i = i / pi With budget constraint: b i = b ( i / Σl  k ) 4/23/2019

Stability and Oscillation Reduction Congestion-sensitive pricing (tatonnement process) has been shown to be stable (see web page for proof) Oscillation reduction Users: re-negotiate only if price change exceeds a given threshold Network: update price only when traffic change exceeds a threshold; negotiate resources in larger blocks between domains 4/23/2019

Test-bed Architecture Demonstrate functionality and performance improvement: blocking rate, average loss and delay, price stability, perceived media quality Host HRN negotiates resources for a system Host processes (HRN, VIC, RAT) communicate through Mbus Network FreeBSD 3.4 + ALTQ 2.2, CBQ extended for DiffServ NRNs: Process RNAP messages Admission control, monitor service statistics, compute price At edge, dynamically configure the conditioners and form charge Inter-entity signaling: RNAP RAT VIC Mbus HRN RNAP NRN 4/23/2019

Functions of Routers Interior routers: per-class policing, e.g, TBMETER (in/out) for a class Edge routers: flow conditioning/policing based on SLA The NRN supports flexible definition of service classes and their specification. The spec is typically based on a set of QoS parameters. It also includes a pricing function for calculating price components. Identifier is typically a standard mechanism to identify the service to clients. For example, for a DiffServ service like Expedited Forwarding, the identifier is the DSCP. The NRN at edge routers performs per-flow policing and conditioning. The policing function can for example use a token-bucket to measure the flow’s usage and take appropriate measures. The NRN's running on interior routers do not maintain per-flow measurements. But they do keep track of per-flow allocation information. They do per-class policing using the DiffServ BA technique. 4/23/2019

Network Resource Negotiator (NRN) Provide price for each service class Measurement-based admission control predict future demand, update congestion price based on predictions 4/23/2019

Network States Per-class bandwidth and price variations blocking reduction due to adaptation 4/23/2019

Adaptive Wireless Terminal WAP development over Nokia Toolkit 2.0 Currently cell phone services: Flat pricing and best effort: all users get worse quality (coarse voice, busy signal, cut off) when congestion occurs. Using our solution: Provide real-time pricing information, e.g., every 10 minutes or every call Customers choices: pay a premium to have best quality pay less by tolerating worse quality back off to call another time. Reduce the blocking rate of overall network 4/23/2019

Simulation Design Performance comparison: Four groups of experiments: Network with dynamic services and rate-adaptive users and network with no adaptive users Fixed price based policy (FP): usage price + holding price versus congestion price based adaptive service (CPA): usage price + holding price + congestion price Four groups of experiments: Effect of traffic burstiness Effect of traffic load Load balance between classes Effect of admission control 4/23/2019

Performance Measures Engineering metrics Economic metrics Bottleneck traffic arrival rate Average packet loss and delay User request blocking probability Economic metrics Average and total user benefit End to end price, and its standard deviation Network revenue 4/23/2019

Simulation Modeling Network Simulator (NS-2) Weighted Round Robin (WRR) scheduler Three classes: EF, AF, BE EF: tail dropping, limited to 50 packets; load threshold 40%, delay bound 2 ms, loss bound 10-6 AF: RED-with-In-Out (RIO), limited to 100 packets; load threshold 60%, delay bound 5 ms, loss bound 10-4 BE: Random Early Detection (RED), limited to 200 packets; load threshold 90%, delay bound 100 ms, loss bound 10-2 Sources: on-off or Pareto on-off (shape parameter: 1.5) Negotiation period: 30 s, session length 10 min 4/23/2019

Simulation Architecture Topology 1 (60 users) Topology 2 (360 users) 4/23/2019

Load Balance between Classes Variation over time of the price of AF class Ratio of AF class traffic migrating through class re-selection 4/23/2019

Effect of Traffic Burstiness Average packet delay Average packet loss 4/23/2019

Price average and standard deviation of AF class Effect of Traffic Burstiness (cont’d) Price average and standard deviation of AF class Average user benefit 4/23/2019

Effect of Traffic Load (cont’d) Average packet loss Average packet delay 4/23/2019

Price average and standard deviation of AF class Effect of Traffic Load Price average and standard deviation of AF class Average user benefit 4/23/2019

Load Balance between Classes (cont’d) Average packet delay Average packet loss 4/23/2019

Effect of Admission Control Average packet delay Average packet loss 4/23/2019

Effect of Admission Control (cont’d.) Average and standard deviation of AF class price User request blocking rate 4/23/2019

Conclusions RNAP Pricing model Application adaptation Supports dynamic service negotiation, mechanisms for price and charge collation, auction bids and results distribution Allows for both centralized and distributed architectures Supports multi-party negotiation: senders, receivers, or both Can be stand-alone, or embedded inside other protocols Reliable and scalable Pricing model Consider both long-term user demand and short-term traffic fluctuation; use congestion-sensitive component to motivate user demand adaptation upon resource scarce Application adaptation Maximize users’ perceptual value, tradeoffs between quality and payment 4/23/2019

Conclusions (cont’d) Auction Model Simulation results Serves more users than comparable schemes, and has less signaling overhead and greater certainty of service availability Simulation results Differentiated service requires different target loads in each class CPA policy coupled with user adaptation effectively limit congestion, provide lower blocking rate, higher user satisfaction and network revenue than with the FP policy Both auction and tatonnement process can be used to calculate the congestion price; auction scheme gains higher perceived user benefit and network utilization at cost of implementation complexity and setup delay Without admission control, service is assured by restricting the load to the targeted level; with admission control, performance bounds can be assured even with FP policy, but CPA reduces the request blocking rate greatly and helps to stabilize price 4/23/2019

Conclusions (cont’d) Future work Allowing service class migration further stabilizes price Users with different demand elasticity share bandwidth proportional to their willingness to pay Even a small proportion user adaptation results in a significant performance improvement for the entire user population Performance of CPA further improves as the network scales and more connections share the resources Future work Propose light-weight resource management protocol Resource management in wireless environment 4/23/2019