LTTCP: EndtoEnd Framework to Improve TCP Performance over Networks with Lossy Channels

1
LT-TCP End-to-End Framework to Improve TCP
Performance over Networks with Lossy Channels

• Student Vijay Subramanian, RPI
• ATT Researchers K. K. Ramakrishnan, Lusheng Ji
• RPI Faculty Shivkumar Kalyanaraman

2
Overall Motivation

• TCP is the dominant transport protocol used in
  the Internet
• TCP has mechanisms designed to avoid congestion
• tolerate loss rates of the order of 1-2 maximum
• Wireless channels becoming more pervasive
• With in-building and access links becoming
  wireless, more than the last hop could be
  wireless.
• Wireless links
• individual links can experience loss that can be
  high in transient situations
• until power and link rate adjustments kick in
• E.g., ad-hoc networks, Mesh networks.
• Interference can also result in high loss rates.
• LT-TCP uses a combination of mechanisms to make
  TCP much more tolerant to loss due to packet
  corruption

3
Sources of Corruption/Loss in Wireless

• Many causes for packet corruption
• Poor signal to noise ratio
• Distance from transmitter walls
• Interference from other transmitters
• Hidden nodes
• Use of the same band by other devices causes
  interference
• E.g., 802.11b and BlueTooth attempt to co-exist
  in the 2.4 Ghz ISM band
• Link layer mechanisms attempt to keep residual
  error rate low
• 802.11b for instance uses rate adaptation to
  reduce packet corruption
• How does that interact with interference?
• Link layer FEC and ARQ mechanisms to overcome
  packet corruption
• With multi-hop wireless networks, overall loss
  rate may still be significant
• end-end schemes such as LT-TCP may play a role in
  overcoming this

4
Problem space

• Dynamic Range
• Can we extend the dynamic range of TCP into high
  loss regimes?
• Can TCP perform close to the residual capacity
  available under high loss rates?
• Congestion Response
• How should TCP respond to notifications due to
  congestion..
• but not respond to packet erasures that do not
  signal congestion?
• Mix of Reliability Mechanisms
• What mechanisms should be used to extend the
  operating point of TCP into loss rates from 0 -
  30 - 50 packet loss rate?
• How can Forward Error Correction (FEC) help?
• How should the FEC be split between sending it
  proactively (insuring the data in anticipation of
  loss) and reactively (sending FEC in response to
  a loss)?
• Timeout Avoidance
• Timeouts Useful as a fall-back mechanism but
  wasteful otherwise especially under high loss
  rates.
• How can we add mechanisms to minimize timeouts?

5
Building Blocks

• ECN-Only We infer congestion solely from ECN
  markings. Window is cut in response to
• ECN signals hosts/routers have to be
  ECN-capable.
• Timeouts The response to a timeout is the same
  as with standard TCP.
• Window Granulation and Adaptive MSS
• We ensure that the window always has at least G
  segments at all times to increase the ACK stream.
• Window size in bytes initially is the same as
  normal SACK TCP.
• Initial segment size is small to accommodate G
  segments.
• Packet size is continually so that we have at
  least G segments. Once we have G segments, packet
  size increases with window size.
• Loss Estimation The receiver continually tracks
  the loss rate and provides a running estimate of
  perceived loss back to the TCP sender through
  ACKs.
• We use an adaptive EWMA approach to estimating
  loss to obtain a smoothed estimate that is biased
  slightly towards higher losses.

6
Building Blocks

• Proactive FEC TCP sender sends data in blocks
  where the block contains K data segments and R
  FEC packets. The amount of FEC protection (K) is
  determined by the current loss estimate.
• Proactive FEC based upon estimate of per-window
  loss rate (Adaptive)
• Reactive FEC Upon receipt of 1 or 2 dupacks,
  Reactive FEC packets are sent based on the
  following criteria.
• Number of Proactive FEC packets already sent.
• Number of holes still left in the block at the
  receiver so that all the data packets in that
  block can be reconstructed.
• Loss rate currently estimated.
• Reactive FEC to complement retransmissions both
  of which will be used to reconstruct DATA packets
  at the receiver.

7
Putting it Together.
Application Data
MSS Adaptation
Granulated Window Size
Window
P-FEC
(n,k)
Window Size
Parameter Estimation
Data
FEC Computation
Loss Estimate
8
Simulation Configuration
9
Performance Results
LT-TCP (Multiple Sources)
10
Link-Layer Interactions

• 802.11b WLANs operate in unlicensed ISM band
• 2.412-2.480 GHz
• Can operate at 11 , 5.5 , 2 or 1 Mbps raw data
  rate.
• Rate adaptation algorithms are typically
  proprietary.
• Many applications are free to operate in this
  spectrum
• Microwave, cordless phones, Wireless USB ,
  Bluetooth
• High probability of interference.
• Rate adaptation may be counter-productive.
• At lower data rates, packets are in the air
  longer and are more vulnerable to narrow-band
  noise such as Bluetooth.
• Operating at higher data rates may be better if
  interference is strong.

11
Interference Scenarios

• Consider a Bluetooth voice application.
• Frequency Hopping over 79MHz spectrum that
  overlaps with WLAN spectrum.
• Voice applications use HV1, HV2 or HV3 data
  encoding that keep the channel busy.
• Strong interference leads to lossy wireless
  channels.
• Simulation Scenarios
• Link-Layer with and without Rate-adaptation
• Transport Layer TCP-SACK or LT-TCP
• Link error rates variable from 0-30 PER

12
Simulation Setup
13
Preliminary Results