A Strategy for Implementing Smart Market Pricing Scheme on DiffServ

1
A Strategy for Implementing Smart Market Pricing
Scheme on Diff-Serv

• Murat Yuksel and Shivkumar Kalyanaraman
• Rensselaer Polytechnic Institute, Troy, NY
• yuksem_at_rpi.edu, shivkuma_at_ecse.rpi.edu

2
Outline

• Literature development
• congestion-sensitive pricing
• the Smart Market (SM) pricing scheme
• Adaptation of SM to diff-serv
• Simulation experiments
• Summary

3
Congestion-Sensitive Pricing

• Increase the price when congestion, decrease when
  no congestion.
• A way of controlling users traffic demand and
  hence, a way of controlling network congestion
• Better resource (bandwidth) allocation
• Fairness
• Problems
• Users dont like price fluctuations!
• Each price change must be fed back to the user
  before it could be applied, i.e. hard to
  implement in a wide area network.

4
The Smart Market (SM)

• Proposed by MacKie-Mason and Varian in 1993
• A congestion-sensitive pricing scheme
• Price-per-packet reflecting congestion costs
• Users make auction by assigning a bid value to
  each packet before sending it into the network.
• The routers maintain a threshold (cutoff) value
  and pass the packets with bids larger than the
  threshold. They give priority to the packets with
  higher bid!
• The cutoff value changes dynamically based on
  local congestion

5
The Smart Market (SM) (contd)

• The price for each packet is the highest cutoff
  value it passed through, i.e. market-clearing
  price.
• Why is SM important?
• The first congestion-sensitive pricing scheme
• Designed for the smallest granularity level (i.e.
  packet) and hence, attempts the highest possible
  congestion-sensitivity for network pricing
• Ideal scheme from an economic perspective because
  of its pure congestion-sensitivity

6
Adaptation to Diff-Serv

• For data plane packets
• Edge routers (ERs)
• write the bid value (b) to the packet header
• and then send the packet into the core
• Interior Routers (IRs)
• maintain a priority queue, sorted according to
  packets bids
• if bltT, drop the packet
• if bgtT, update the packets clearing-price field
  and forward it
• For control plane packets
• ERs and IRs maintain a time interval (t) which is
  greater than round-trip time (RTT) to operate.
• Hence, the customers are fed back with the
  current price and their account information at
  every t.

7
Adaptation to Diff-Serv (contd)

• ERs and customers
• Ingress-ER sends a probe packet to the network
  core at every t to find out the current
  clearing-price of the network.
• Egress-ER responds to the probe packet by a
  feedback packet that includes current
  clearing-price and bill to the customer.
• set the bids of control packets to the maximum
  bid value (limitation-- bids must be bound to a
  range)
• Ingress-ER informs the customer about his bill
  and the current clearing-price.
• Customers adjust their bids and traffic based
  upon the bill, the clearing-price, and their
  utility.
• IRs
• update the threshold (T) value at every t
• update control packets clearing-price field too

8
Cutoff Value, T

• SM says that the IRs should adjust the cutoff
  value such that T n/K D(Y), where n is the
  number of customers and K is the capacity of the
  network.
• IRs update T by calculating D(Y) at the end of
  each interval, t.
• We used the following approximation for
  calculating T
• where Di is the average delay at interval i,
  and Ti is the cutoff value for interval i.

9
Simulation Experiments

• Packet size is 1000bytes.
• Propagation delay is 0.1ms on bottleneck links
  and 10ms on the others.
• RTT is 24ms.
• The time interval t is 1000ms.
• User utility is concave u(x) w log(x)
• Users have a budget w and maximize their surplus
  by sending at a rate w/p.
• We simulated two versions of SM
• SM-SORTED higher bids have priority at IRs
• SM-FIFO first-come first served

10
Simulation Experiments (contd)

• 3 user flows with budgets 100, 75 and 25 /Mb.
• Total simulation time is 3000s.

11
Simulation Experiments (contd)
12
Simulation Experiments (contd)
13
Simulation Experiments (contd)

• To observe service differentiation
• Two flows with a varying ratio of budgets.

14
Simulation Experiments (contd)

• Each user flow has a budget of 10/Mb.

15
Simulation Experiments (contd)
16
Summary

• Major changes to SM are need for an
  implementation on diff-serv
• By extensive simulation we observed that
• SM can control congestion with low queues and
  high utilization
• Packet sorting (i.e. priority to higher bids)
  degrades system performance
• SM performs in between max-min and proportional
  fairness