Presentation is loading. Please wait.

Presentation is loading. Please wait.

U NIVERSITY OF M ASSACHUSETTS, A MHERST Department of Computer Science Towards a Robust Protocol Stack for Diverse Wireless Networks Arun Venkataramani.

Similar presentations


Presentation on theme: "U NIVERSITY OF M ASSACHUSETTS, A MHERST Department of Computer Science Towards a Robust Protocol Stack for Diverse Wireless Networks Arun Venkataramani."— Presentation transcript:

1 U NIVERSITY OF M ASSACHUSETTS, A MHERST Department of Computer Science Towards a Robust Protocol Stack for Diverse Wireless Networks Arun Venkataramani (in collaboration with Ming Li, Devesh Agrawal, Deepak Ganesan, Aruna Balasubramanian, Brian Levine, Xiaozheng Tie at UMass Amherst)

2 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 2 Wireless network landscape today Diverse wireless networks coexist Adapt TCP/IP stack in different ways WLAN Mesh MANET DTN Little mobility High mobility Some mobility Cellular

3 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 3 Interconnecting diverse networks? Vision: A simple robust protocol stack to interconnect diverse wireless networks. DTN WLAN Mesh MANET Little mobility High mobility Some mobility Cellular

4 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 4 Why does TCP/IP not suffice? E2E route disruptions cause unavailability DTN WLAN Mesh MANET Little mobility High mobility Some mobility Cellular

5 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 5 Why does TCP/IP not suffice? Gap between what you buy vs. get DTN WLAN Mesh MANET Little mobility High mobility Some mobility Cellular Advertised capacity: 11Mbps

6 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 6 Principle: Design for high uncertainty Observe what works at an extreme point in the design space--always-partitioned DTNs--for insights into a robust design.

7 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 7 Outline Motivation Block transport Replication routing Research challenges

8 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 8 E2E Feedback 1. E2E Transport E2E rate control is error-prone E2E retransmissions are wasteful Rate Control Rate Control Dest Source Congestion? Link loss? Loss Rate RTT

9 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 9 1. E2E Transport E2E rate control is error-prone E2E retransmissions are wasteful Source Dest Redundant Transmissions E2E Retransmissions P

10 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 10 1. E2E Transport E2E rate control is error-prone E2E retransmissions are wasteful E2E route disruptions cause unavailability Route disruptions due to mobility

11 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 11 2. Packet as Unit of Control Channel access Link layer ARQ Listen Backoff RTS/CTS

12 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 12 2. Packet as Unit of Control Channel access Link layer ARQ Timeout Backoff Timeout Backoff

13 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 13 3. Complex Cross-Layer Interaction Transport Link Link-layer ARQ/backoff hurts TCP rate control Highly Variable RTT Link ARQ Rate Control Rate Control

14 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 14 Hop: Clean-Slate Transport Protocol End-To-End Packets Complexity Hop-by-Hop Blocks Minimize interaction Block-switched networks: A new paradigm for wireless networks, Li et al. [NSDI 2009]

15 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 15 Hop Design Reliable Block Transfer ACK Withholding Micro-block Prioritization Virtual Retransmission Backpressure Per-hop Multi-hop

16 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 16 Reliable Per-Hop Block Transfer B-SYN Request for B-ACK B-ACK Packet bitmap Mechanisms Burst mode (TXOP) Block ACK based ARQ Benefits Amortizes control overhead CSMA TXOP

17 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 17 Hop Design Reliable Block Transfer ACK Withholding Micro-block Prioritization Virtual Retransmission Backpressure Per-hop Multi-hop

18 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 18 Mechanism Leverages in-network caching Re-transmits blocks only when unavailable in cache Benefits Fewer transmissions Low overhead Simple Virtual Retransmission (VTX) B-SYN B-ACK A B E C D DATA

19 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 19 Mechanism Leverages in-network caching Re-transmits blocks only when unavailable in cache Benefits Fewer transmissions Low overhead Simple Virtual Retransmission (VTX) A B ED B-SYN VTX B-ACK B-SYN VTX B-SYN with VTX flag set VTX Timer E2E ACK B-SYN VTX Timer

20 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 20 Hop Design Reliable Block Transfer ACK Withholding Micro-block Prioritization Virtual Retransmission Backpressure Per-hop Multi-hop

21 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 21 Mechanism Limits outstanding blocks per-flow at forwarder Backpressure ABCDE Source Dest Slow Limit on outstanding blocks=2 B-SYN Hold B-ACK B-SYN

22 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 22 Mechanism Limits outstanding blocks per-flow at forwarder Benefits Improves network utilization Backpressure ABC D E Source Dest F G Aggregate goodput without backpressure: 6Mbps 20Mbps 10Mbps 20Mbps 1Mbps

23 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 23 Mechanism Limits #outstanding blocks per-flow at forwarder Benefits Improves network utilization Backpressure A BC DE Source Dest F G Aggregate goodput with backpressure: 10Mbps 20Mbps 10Mbps 20Mbps 1Mbps Limit of Outstanding Blocks=1

24 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 24 Hop Design Reliable Block Transfer ACK Withholding Micro-block Prioritization Virtual Retransmission Backpressure Per-hop Multi-hop

25 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 25 Mechanism : Receiver withholds all but one BACK Benefit: Low overhead Less conservative Simple Ack Withholding ACB B-SYN B-ACK DATA B-SYN B-ACK DATA Withhold B-ACK

26 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 26 Hop Design Reliable Block Transfer ACK Withholding Micro-block Prioritization Virtual Retransmission Backpressure Per-hop Multi-hop

27 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 27 Micro-block Prioritization Mechanisms Sender piggybacks small blocks to B-SYN Receiver prioritizes small block’s B-ACK Benefits Low delay for small blocks SSH FTP B-SYN B-ACK DATA B-SYN 64B 1MB Sender Receiver

28 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 28 20 nodes on the 2 nd floor of UMass CS building Apple Mac Mini 1.8GHz, 2GB RAM, Atheros 802.11 a/b/g card Testbed

29 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 29 Single-flow Single-hop Performance Hop achieves significant gains over TCP HopTCP 28x 1.6x 1.2x

30 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 30 Single-flow Multi-hop Performance HopTCP 2.7x 2.3x 1.9x Hop achieves significant gains over TCP

31 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 31 Graceful Degradation with Loss Emulated link layer losses at the receiver TCP breaks down at moderate loss rates HopTCP

32 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 32 Scalability to High Load 30 concurrent flows Hop is much fairer than TCP Mean goodput HopTCP Hop-by-hop TCP 150x 20x 2x

33 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 33 Hop Performance Breakdown

34 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 34 Hop over WLAN Hop improves utilization over TCP+RTS/CTS AP Mean (kbps) Median (kbps) Hop663652 TCP587244 TCP+RTS/CTS463333

35 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 35 Low Delay for Small Transfers AP 1 small transfer competing with 4 large Small transfer size (KB) Hop lowers delay across all file sizes

36 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 36 Impact on Concurrent VoIP

37 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 37 Hop in Partitionable Networks ProtocolGoodput (Kbps) Hop (H=1)320 (±29) Hop (H=100)457 (±18) DTN2.5159 (±15) Hop significantly outperforms (TCP-based) DTN2.5

38 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 38 Network and Link Layer Dynamics 30 concurrent flows with bit rate control +/ OLSR Hop significantly outperforms TCP under dynamic network conditions

39 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 39 Hop with 802.11g Hop continues to outperform TCP with 11g

40 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 40 Related Work Fixing E2E rate-control [Snoop, WTCP, Westwood+, ATCP, TCP-ELFN] Separating loss/congestion [Snoop, WTCP, Westwood+, ATCP, TCP-ELFN] [ATP, NRED, IFRC, WCP] Network-assisted rate control [ATP, NRED, IFRC, WCP] Hop circumvents rate control Backpressure [Tassiulas et al.] ATM, theoretical work [Tassiulas et al.] [Fusion, Flush] Tree/chain sensor data aggregation [Fusion, Flush] [RAIN, CXCC, Horizon] Reliable point-to-point transport [RAIN, CXCC, Horizon] Hop reduces backpressure overhead using blocks Batching [802.11e/802.11n, WiLDNet, Kim08, CMAP][Delayed-ACK, DTN2.5][ExOR] Common optimization at link [802.11e/802.11n, WiLDNet, Kim08, CMAP], transport [Delayed-ACK, DTN2.5], and network [ExOR] layers Hop leverages batching across layers

41 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 41 Outline Motivation Block transport Replication routing Research challenges

42 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 42 Replication routing P XZY P P W P Replication reduces delay under topology uncertainty

43 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 43 Routing as resource allocation problem RAPID Protocol (X,Y): 1. Control channel: Exchange metadata 2. Direct Delivery: Deliver packets destined to each other 3. Replication: Replicate in decreasing order of marginal utility 4. Termination: Until all packets replicated or nodes out of range YX Change in utility Packet size Problem : Which packets to replicate given limited bandwidth to optimize a specified metric? DTN routing as a resource allocation problem, Balasubramanian et al. [SIGCOMM 2007]

44 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 44 Objective: Minimize average delay Define U(i) = -(T + D) T = time since created, D = expected remaining time to deliver Simplistic assumptions uniform exponential meeting with mean ¸ global view Utility computation example D = ¸ i X D = ¸/2 i Y D =¸/3 i W Z

45 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 45 Utility computation example (cont’d) Replicate i Metric: Min average delay Deadline of i < T Deadline of j = T 1 > T Metric: Maximize #packets delivered within deadline Replicate j i X j Y j WZ

46 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 46 Deployment on DieselNet

47 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 47 Results based on DieselNet traces

48 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 48 Outline Motivation Block transport Replication routing Research challenges

49 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 49 C1: When to replicate? Replication useful under Topology uncertainty Delay optimization Light load scenarios WLAN Mesh MANET DTN Low uncertainty Forwarding suffices High uncertainty Replication useful X1X1 X2X2 XkXk src dst relays X i = r.v. for delay of path i Forwarding delay = min i (E[X i ]) Replication delay = E[min i (X i )] Claim: Delay benefit of replication is unbounded.

50 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 50 C2: Bit Rate Control with Blocks  Reduces negative interaction between bit rate control and Transport ARQ  Simplifies collision vs. channel loss detection  Using small frames  Reducing collisions Find r: max(r(1-p(r))) ARQ => keep p(r) low TCP => keep p(r) low

51 U NIVERSITY OF M ASSACHUSETTS A MHERST Department of Computer Science 51 Conclusions Vision: Simple robust protocol stack for diverse wireless networks Key building blocks Block transport (Hop) Replication routing (RAPID) Source code http://hop.cs.umass.edu http://prisms.cs.umass.edu/rapid


Download ppt "U NIVERSITY OF M ASSACHUSETTS, A MHERST Department of Computer Science Towards a Robust Protocol Stack for Diverse Wireless Networks Arun Venkataramani."

Similar presentations


Ads by Google