CMT-QA：无线异构网络中质量感知的自适应多路径传输

需积分: 10 7 浏览量更新于2024-08-12 收藏 1.88MB PDF 举报

"CMT-QA：异构无线网络中质量感知的自适应并行多路径数据传输" 在当前的移动通信环境中，设备通常配备多个网络接口，以利用流控制传输协议（SCTP）提供的多路径并行传输功能，从而提升吞吐量和带宽聚合。然而，由于不同路径的带宽和延迟差异，数据可能会出现乱序接收，这可能导致严重应用程序性能下降。针对这一问题，研究者提出了CMT-QA（质量感知自适应并发多路径传输）方案，它专门用于无线异构网络中的FTP式数据传输和实时视频服务。 CMT-QA的核心在于定期监控和分析每条传输路径的数据处理能力，进而做出适应性的数据传递决策，选择最佳路径进行并行数据传输。这一解决方案包含一系列机制，能够智能地在多个路径之间分配数据块，并独立控制每个路径的数据流量速率。CMT-QA致力于减少因路径差异导致的数据重排序延迟，以及不必要的快速重传，以减轻乱序接收对系统性能的影响。此外，CMT-QA能区分不同类型的丢包情况，避免因不合理的拥塞窗口调整而频繁进行重传。这有助于优化网络资源的使用，提高传输效率。通过这种方法，CMT-QA不仅提升了数据传输的速度，还确保了服务质量的稳定性。仿真实验结果显示，CMT-QA在性能和服务质量方面显著优于传统的多路径传输解决方案。这意味着CMT-QA对于提升无线环境下的多媒体传输质量和用户体验具有重要意义。这一研究论文的贡献在于提供了一个创新的、基于SCTP的传输策略，有望在未来的无线通信系统中得到广泛应用，特别是在高带宽需求和复杂网络环境下。 CMT-QA是针对异构无线网络中数据传输挑战的一种有效应对策略，它通过质量感知和自适应性来优化多路径传输，减少性能损失，提高数据传输的可靠性和效率。这项技术的应用将有助于推动移动设备在多种网络条件下的高效数据通信，为用户带来更好的通信体验。

the paths diversity by taking potential path correlation

into account, which avoids underlying shared bottle-

necks. A rate allocation model for best path transfer

was presented by our team in [19] and we showed

that it achieves the global optimum. However, none of

the above works consider the dynamic path selection

according to the likely variation of the current network

conditions.

Yilmaz

et al.

[20] introduced non-renegable selective

acknowledgements (NR-SACK) in order to avoid retain-

ing the non-outstanding gap ACK chunks in the sender

buffer. NR-SACK gives possibility to free buffer space

earlier and reuse it for new data chunks. Dreibholz

et al.

[21] presented a blocking fraction factor and proposed

a preventive retransmission policy based on the factor

for effective transmission. Adhari

et al.

[22] proposed

an optimized strategy to enhance the send and receive

buffer handling by avoiding one path to dominate the

buffer occupation. These solutions achieve performance

improvements. However, the researchers do not provide

a proper data distribution mechanism to ensure data

packets arrival at receiver in order as much as possible.

Cui

et al.

[23] introduced a fast selective ACK scheme

for SCTP to enhance transmission throughput in mul-

tihoming scenarios. In the networks with asymmetric

delays for forward and reverse paths, a multihomed

receiver sends SACK chunks to the sender over the

fastest reverse path, which facilitates to inﬂate the con-

gestion window and to retransmit the lost data packets

as quickly as possible. Yet the solution just considers the

transmission of control chunks and fails to enhance the

overall transmission efﬁciency.

3 CMT-QA SYSTEM DESIGN OVERVIEW

During multihomed communications in a heterogeneous

wireless networks, delay, bandwidth and loss rate of al-

ternative paths can be signiﬁcantly different. If a round-

robin data delivery approach is used, slower paths are

easily overloaded, while faster paths remain underuti-

lized. In order to avoid unbalanced transmissions, re-

duce received data reordering and alleviate the receiver

buffer blocking problem caused by the use of dissimilar

paths using CMT, CMT-QA makes important contribu-

tions in the following three stages:

• Accurately senses each path’s current transmission

status and estimates in real time each path’s data

handling capacity.

• Includes a newly designed data distribution algo-

rithm to deliver optimally the application layer data

over multiple paths and ensure the received data

arrives in order.

• Introduces a proper retransmission mechanism to

handle different kinds of packet loss and alleviate

the packet reordering problem.

Fig. 2 illustrates the design of the CMT-QA architec-

ture, which includes a Sender, a Receiver and n commu-

nications Paths via the heterogeneous wireless network

sender

buffer

Stream 1

Stream 2

Stream N

Data

Distribution

Scheduler

sender

Path 1

Stream 1

Stream 2

Stream N

Receiver

network

Path status feedback information

Path 1

Path condition

Optimal

Retransmission Policy

Calculate the

estimation interval

Path handling

capability estimater

Sender

Reassembly

in receiver

buffer

Unbound

and

collect

data

chunks

Path 2

Path n

Path 2

Path n

Path Quality

Estimation Model

Fig. 2. CMT-QA architecture.

environment. The Receiver receives data and recreates

the original data chunks, if multiple data and control

chunks are bundled together by the Sender into a single

SCTP packet for transmission. In the case in which a

user message is fragmented into multiple chunks, the

Receiver reassembles the fragmented message in the re-

ceiver buffer before its delivery to the user. The feedback

information of path status in the network is collected by

the Sender and used to estimate the path quality. At the

Sender there are three major CMT-QA blocks which are

the

Path Quality Estimation Model

(PQEM),

Data Dis-

tribution Scheduler

(DDS) and

Optimal Retransmission

Policy

(ORP). CMT-QA aims to intelligently adjust data

distribution for each path and support in order data

packet arrival at destination.

PQEM chooses a reasonable estimation interval to

calculate the data handling rate of entering and leaving

sender buffer for each path, which describes any path’s

communication quality. Any unfavorable conditions in-

cluding packet loss rate, link delay, buffer size of routers,

channel capacity and number of other data ﬂows etc.

will determine performance degradations of the paths

handling capability. PQEM uses a comprehensive eval-

uation method to reﬂect the impact of above factors

on the communication quality. PQEM’s data handling

rate of sender buffer describes better the end-to-end

delivery conditions as its shorter estimation time enables

its timely reﬂection of the current communication path

status. Additionally, the samples for the time interval of

distributed data’s entering and leaving the sender buffer

can be obtained easily to predict the path quality change

trends.

Based on the path quality estimation results by PQEM,

DDS chooses a subset of suitable paths for load sharing

and dynamically assigns them appropriate data ﬂows.

In meanwhile, by forecasting the time of data arriving at

destination in terms of each path quality, DDS can draft

the period of packet distribution and also can adjust

the distributed data amount for each path. In this way,

the application data chunks are intelligently dispatched

over multiple paths in real-time. Compared with the

round robin scheme, we believe that the most effective

approach to mitigate the reordering is to use a heuristic

mechanism to decide the fraction of data scheduled to

be transmitted on each path. The data distribution rate

剩余13页未读，继续阅读

weixin_38679233

粉丝: 2
资源: 872

CMT-QA：无线异构网络中质量感知的自适应多路径传输

基于自适应网络编码的异构无线链路并发传输控制方法研究.docx

CMT-SC: 异构无线环境下的CMT丢包识别与优化

matlab代码abs-CMT-CNN:学习用于鲁棒行人检测的跨模态深度表示

intel-cmt-cat：用于英特尔（R）Resource Director技术的用户空间软件

yt-cmt-analyzer:一个用于对YouTube评论进行情感分析的Web应用程序

CMT-NC: Improving the Concurrent Multipath Transfer Performance using Network Coding in Wireless Networks

cmt-js:CMGTools JS是一个JavaScript库，可为CMSGears提供实用程序，UI组件和MVC框架实现

CMT4A-CMSS:社会科学的计算方法

holiday-card-cmt:由Classroom为GitHub创建的holiday-card-cmt

异构无线网络中用于移动视频流的失真感知并发多路径传输

最新资源