WP2 – Network and interaction

Objectives

  • Explore latency issues for future infrastructures
  • Acquire a deep understanding of the latency/throughput tradeoff from a network perspective
  • Make the network routers more suitable for low latency traffic
  • Explore the use of alternative queue management for low latency, bearing in mind scalability issues and the massive amount of flows in the next generation networks
  • Investigate Internet deployment of combined network/end system solutions
  • Develop methods to make home gateways and wireless access points appropriate for low-latency traffic
  • Develop robust methods for end-system/home gateway interaction to enable low-latency communication

Description of work

Excessive buffering in the network can be a large contributor to high end-to-end latencies. In response, WP2 investigates network techniques for reducing latency including network techniques that work in interaction with end systems. The WP is structured into 4 tasks. Task 2.1 analyzes how various network issues influence latency and how they interact with the transport layer. The result of the analysis will be used as input for the design and implementation work in tasks 2.2. and 2.3, where task 2.2 focuses on network techniques that do not require transport layer support and task 2.3 focuses on network techniques that work in interaction with end systems. Tasks 2.2 and 2.3 will start a few months into the project when initial results from the analysis in task 2.1 can be used to guide the work. Task 2.4 will start from the beginning of the project and aims to bring the gained knowledge and developed network mechanisms to the IETF for presentation and evaluation. This will provide important feedback for the work and help maximize the impact as results on Internet transport need to be standardized to have global impact.

Task 2.1: Analysis of network- and buffer issues, and study of network/transport protocol interactions

Description: This task focuses on identifying and analysing the network issues that today influence the delay experienced by traffic on the Internet. The interactions between network and transport layer will be analyzed to identify the root causes that impact the delay experienced by time-sensitive traffic in the Internet. The work will include examining the architectural constraints of present systems and identifying opportunities to change the architecture in future networks.

  • We will regard how different deployed TCP flavors interact in network buffers with latency- sensitive flows.
  • We will also analyse the relationship between network buffer sizes in access links (which have an impact on delay) and the behaviour of TCP congestion control variants.
  • Analysis of where the causes for delay are located, within the broadband access and aggregation network, will be performed in this task. One of the focus points will be the delay introduced by the broadband access and aggregation router. Traditional queue management techniques that are used today (tail drop, RED, etc.) on xDSL and FTTx access nodes will be assessed with respect to low-latency requirements.

Task 2.2: Development of network techniques

Description: This task deals with techniques that will be designed for reducing the delay experienced by low-latency traffic. This work will focus on changes to the way routers buffer and forward Internet packets, acting independently of the transport protocol mechanisms used. The optimisations include solutions for reducing network delay experienced by low-latency traffic, with emphasis on access networks and home gateways.

  • We will investigate techniques applicable in access networks and home gateways including home gateway WLANs and NATed networks. These techniques include new mechanisms for reducing network buffer-induced latency, e.g. by using AQM techniques and virtual queues.
  • We will design and evaluate novel ways to manage the queues in the access and aggregation router to reduce the delay experienced by delay-sensitive applications while at the same time enforcing fairness between delay sensitive and insensitive connections. These techniques will be studied for both the traditional transport protocols and for the new transport level mecha- nisms developed in the project.

This task will be performed in two sequential phases: 1) Simulations and initial development of mechanisms.  2) Prototype development of mechanisms ready for deployment and evaluation of the developed mechanisms.

Task 2.3: Development of network techniques including end-system interaction

Description: This task concentrates on the design and assessment of network techniques that require support from transport protocols implemented in end systems to get the full benefit of the optimisations.

  • As a concrete starting point, we will explore how to deploy AQM-based congestion control approaches similar to DCTCP over the Internet. The behaviour of DCTCP-like approaches in non-controlled networking environments will be evaluated, as well as new techniques for detecting queue build-up from end systems. Next generation techniques based on the DCTCP paradigm for tightly-controlled environments with low latency, like e.g. intra data-center com- munication, will be explored as well.
  • By studying how bulk transfer streams sometimes override time-dependent streams, we will learn how to create systems that allow for time-dependent traffic to coexist with bulk traffic while keeping the latency as low as possible. Mechanisms for improving such interaction will be developed both for systems where all streams share one interface, and for systems using multiple interfaces. We will evaluate such mechanisms in the context of optimisations developed in WP1 (task 1.2).
  • For an application, the host may share information with a home gateway that will help reduce latency for selected traffic. Approaches involving resource sharing between end systems and home gateways will be investigated and supported by testbed experimentation. We will design and evaluate lightweight and scalable information signaling from the network elements, with special attention for interactive video services.

This task will be performed in two sequential phases: 1) Simulations and initial development of mechanisms. 2) Prototype development of mechanisms ready for deployment and evaluation of the developed mechanisms.

Task 2.4: IETF presentation and evaluation of network results from RITE

Description: By presenting the results from RITE to the IRTF/IETF experts, we will gain valuable feedback and identify potential problems. Our group of IRTF/IETF experts will work with the stan- dardization body to create drafts, and ultimately standards based on RITE work. Such standards will help the industry in implementing the RITE results into existing systems, allowing lower latency for a wide range of users. The feedback from IRTF/IETF discussions will loop back into the work on the appropriate task, and help to improve the quality of the final research results.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s