课题基金 / 基金详情

Efficient Marking Techniques for Differentiated Services in the Internet

Efficient Marking Techniques for Differentiated Services in the Internet
互联网差异化服务的高效标记技术
批准号:
0087548
负责人:
Prasant Mohapatra
金额:
$17.53万
依托单位:
依托单位国家:
美国
项目类别:
Standard Grant
财政年份:
2001
资助国家:
美国
项目状态:
已结题
起止时间:
2001-09-15 至 2001-11-30

项目摘要

项目成果

Prasant Mohapatra的其他基金

相似基金

相关文献

中文摘要
翻译
当前一代互联网的尽力而为服务模式不足以满足不断发展和扩展的应用,例如连续媒体、实时应用和电子商务环境。这些应用需要一定的服务质量(Qos)保证,这可以通过互联网中的服务区分来提供。区分服务(DiffServ)已经被提出作为在因特网中提供服务区分的可行解决方案。DiffServ在每跳的基础上对聚合流量进行操作,从而避免了与集成服务模型相关联的可扩展性问题。DiffServ体系结构的一个主要组件是基于相邻因特网域(域间制造商)之间或终端节点与其第一个域之间的服务级别协议(叶标记)来标记分组的标记。互联网核心路由器根据它们的标记区别地路由分组。因此,标记技术对区分服务的性能有很大的影响。该方案的重点是设计和实现高效的互联网区分服务标记技术。所提出的研究针对域间标记和叶标记的标记问题。我们正在进行的域间标记标记方案解决了数据包的随机和早期降级和升级问题,以提高带宽利用率,同时保持不同流之间的公平性。该方法采用三色标记方案,并使用改进的漏桶技术进行流量控制。将使用ns模拟器调查详细的性能研究、各种参数的影响、多优先级的使用以及流量控制技术。在叶标记的情况下,我们提出在末端节点将标记方案与TCP层相结合。因此,除了域间层的聚合级流量控制外,我们还将实现某种程度的流级服务差异化。我们的方法是基于所需的服务级别在逻辑上分割TCP拥塞窗口。在拥塞期间,拥塞窗口的分割受到对应于它们的相对服务级别协议的影响。这样的方法还将使具有不同服务级别协议的多个流之间的流控制公平。在以前的几个研究中,维持比例公平一直是一个困难的任务。我们建议通过一个实验实现来评估所提出的叶标记的影响。拟议研究的结果将对互联网基础设施及其用户产生更广泛的影响。具体地说,所提出的技术将在提供区分服务的同时提高互联网的带宽利用率。在地方层面,这项研究将对密歇根州立大学的学生产生积极的教育影响。
英文摘要
The best-effort service model of the current generation Internet is not adequate for evolving and expanding applications such as continuous media, real-time applications, and e-commerce environments. These applications require certain quality of service (QoS) assurance that could be provided through service discrimination in the Internet. Differentiated Services (DiffServ) have been proposed as a viable solution for providing service differentiation in the Internet. DiffServ operates on the aggregate traffic on a per-hop basis and thus avoids the scalability problems associated with the integrated service model. A major component of the DiffServ architecture is the markers that mark packets based on the service level agreements between adjacent Internet domains (interdomain makers) or between the end node and its ver first domain (leaf markers). The Internet core routers route packets discriminatley based on their marking. Thus the marking technique affects the performance of DiffServ significantly. This proposal focusses on the design and implementation of efficient marking techniques for differentiated services in the Internet. The proposed research targets the marking issues at both the interdomain markers and the leaf markers. Our ongoing effort on the marking scheme for interdomain marker addresses random and early demotion and promotion of packets to improve the bandwidth utilization while maintaining fairness between different flows. This approach employs a three-color marking scheme and uses a modified leaky bucket technique for flow control. A detailed performance study, impact of various parameters, usage of multiple precedence levels, and flow control techniques will be investigated using the ns simulator. In the context of leaf markers, we propose to integrate the marking scheme with the TCP layer at the end nodes. Thus in addition to the aggregte level flow control at the interdomain layer, we will also achieve some degree of flow-level service differentiation. Our approach is based on splitting the TCP congestion window logically on the basis of the desired levels of services. During congestion, the splits of the congestion window gets affected corresponding to their relative service level agreements. Such an approach would also make the flow control fair across multiple flows with different service level agreements. Maintenance of proportional fairness has been observed as a difficult task in several prior research. We propose to evaluate the impact of the proposed leaf marker through an experimental implementation. The outcome of the proposed research will have a broader impact on the Internet infrastructure and on its users. Specifically, the proposed techniques will improve the bandwidth utilization in the Internet while providing differentiated services. At the local level, the research will have a positive educational impact on the students of Michigan State University.
期刊论文(0)
专著(0)
科研奖励(0)
会议论文
CNS: Collaborative Research: EAGER: From Computer Networks to Food Networks
  • 批准号:
    1414464
  • 项目类别:
    Standard Grant
  • 资助金额:
    $7.5万
  • 财政年份:
    2014
  • 负责人:
    Prasant Mohapatra
  • 依托单位:
NeTS: Small: Collaborative Research: Towards a User Centric Battery Management System for Smartphones
  • 批准号:
    1319721
  • 项目类别:
    Standard Grant
  • 资助金额:
    $20.0万
  • 财政年份:
    2013
  • 负责人:
    Prasant Mohapatra
  • 依托单位:
EAGER: Deep Network Analysis
  • 批准号:
    1251029
  • 项目类别:
    Standard Grant
  • 资助金额:
    $15.0万
  • 财政年份:
    2012
  • 负责人:
    Prasant Mohapatra
  • 依托单位:
Graduate Research Fellowship Program (GRFP)
  • 批准号:
    1148897
  • 项目类别:
    Fellowship Award
  • 资助金额:
    $94.31万
  • 财政年份:
    2011
  • 负责人:
    Prasant Mohapatra
  • 依托单位:
海外基金