Multiplex communications – Data flow congestion prevention or control – Control of data admission to the network
Reexamination Certificate
2000-08-10
2002-08-06
Nguyen, Chau (Department: 2663)
Multiplex communications
Data flow congestion prevention or control
Control of data admission to the network
C370S395430
Reexamination Certificate
active
06430154
ABSTRACT:
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
N/A
BACKGROUND OF THE INVENTION
The present invention relates generally to computer network and telecommunication technologies, and more specifically to a system and method for supporting multiple application traffic types over a connection based network.
Data associated with different types of applications must often be transferred over a shared communications network. For example, there is a growing need to use a network to transfer data for both what may be referred to as “inelastic” applications, as well as data for what may be referred to as “elastic” applications. “Inelastic” applications are those which cannot tolerate a significant delay variation within the network. Examples of inelastic applications are real time video and/or voice applications. Elastic applications, in contrast, are those which are relatively tolerant of delay variation within the network. Examples of elastic applications include electronic mail and file transfer. In many situations, a customer may desire to use a single network to transfer data associated with both inelastic and elastic applications.
Communications networks are sometimes described as being either “connection-oriented”, or “connectionless.” For example, the classic model of IP (Internet Protocol) routing is connectionless. A connectionless network such as classic IP is based on packets that are self describing, so that each packet can be treated independently. Accordingly, different packets may take different routes between the same two systems. In contrast, connection-oriented networks establish a fixed route between source and destination, over which all traffic is communicated. Examples of connection oriented network technologies include Asynchronous Transfer Mode (ATM) and Multiprotocol Label Switching (MPLS).
When both inelastic traffic and elastic traffic are to be transmitted over a connection oriented network, a decision may need to be made as to what type or quality of connection to use. For example, some connection types may be more appropriate for inelastic traffic than for elastic traffic, and vice versa. Such a scenario arises, for example, in the case of an end station having an interface to a connection oriented network. When the end station emits data units onto the connection-oriented network, it may need to choose from among connections associated with various types of services in order to define the characteristics of the connection or connections to use for either or both of the inelastic and/or elastic traffic.
A similar problem may arise when mapping services provided over a connectionless network to services offered over a connection-oriented network. A specific example of such a scenario is mapping data units from a traditional IP network onto an ATM network. In this case, the services offered by the IP network may include those based on the IP Differentiated Services model. The services offered by the ATM network do not directly map to the IP Differentiated Services model. This arises from the fact that while ATM Quality of Service (QoS) parameters are absolute, i.e. the service offered to one ATM connection is independent of the service offered to another connection, IP Differentiated Services supports service definitions that are relative, such that the QoS available to one flow depends on that offered to another flow. In addition, the notion of a Per-Hop Behavior is fundamental to the Differentiated Services model. However, a Per-Hop Behavior is fundamentally different from an ATM service category. ATM service categories define services whose performance is specified on an end-to-end basis, while Per-Hop Behaviors describe outcomes observable at individual switches.
More specifically, since IP has traditionally been a “best effort” service, QoS like that provided by ATM networks has been difficult to provide over IP. QOS features would allow ISPs to provide different service levels to different customers. IP Differentiated Services is one mechanism intended to provide QoS-like features over IP.
IP Differentiated Services (also referred to as “DiffServ,” or “DS”) is a protocol for specifying and controlling network traffic so that certain types of traffic get precedence. For example, voice traffic, which requires a relatively uninterrupted flow of data, might get precedence over other kinds of traffic. In the IP Differentiated Services model, each IP packet is associated with certain forwarding behaviors—known as Per Hop Behaviors (PHBs). A six-bit value stored in a Differentiated Services (DS) field of the Internet Protocol (IP) header is used to store a Differentiated Services Code Point (DSCP), which specifies the per hop behavior. In particular, two per-hop behaviors, Expedited Forwarding (EF) and Assured Forwarding (AF) have been defined. The EF PHB is intended for applications which are based on “inelastic” traffic, such as those that require a “virtual leased line” service, with associated low loss and low delay variation service commitments. The AF PHB is regarded as suitable for applications based on “elastic” traffic, which are more tolerant of delay variation. In this description, references to “inelastic” traffic are intended to indicate IP traffic associated with the Expedited Forwarding PHB, while references to “elastic” traffic are intended to indicate IP traffic associated with the Assured Forwarding PHB.
ATM provides connections that define routes to be used between communicating entities. ATM supports different types of service that may be associated with a connection. The various service categories supported by ATM enable ATM networks to provide what are generally referred to as Quality of Service (QoS) commitments to users of ATM connections. The set of ATM service categories includes Constant Bit Rate (CBR), and Unspecified Bit Rate (UBR) The CBR ATM service category is associated with connections that provide a fixed (static) amount of bandwidth. A CBR connection is characterized by a Peak Cell Rate (PCR) value that is continuously available during the connection lifetime. The connection source may continuously transmit cells at or below the PCR at any time, and for any duration, with the expectation of low loss delivery. The CBR service provides a tightly constrained Cell Delay Variation (CDV). In contrast, the UBR ATM service is a “best effort” service, intended for applications which do not require tightly constrained delay variation, nor a specified quality of service. UBR service supports a high degree of statistical multiplexing across sources, also referred to as a high “multiplexing gain,” resulting in improved utilization of network resources. UBR service does not provide traffic related service commitments. Specifically, UBR does not include the notion of a per-connection bandwidth commitment.
The International Telecommunication Union has defined a number of ATM-layer Transfer Capabilities (ATCs) corresponding to the ATM Service Categories. For example, the ITU ATC corresponding to the CBR ATM service category is referred to as DBR (Deterministic Bit Rate) with QoS class
1
, and may be written as DBR.
1
. Similarly, the ATC corresponding to the UBR ATM service category is DBR with an unspecified QoS class, written as DBR.U. In the present disclosure, references to ATM services are used for purposes of illustration, but are intended to indicate the respective ITU ATC designations as well.
Various efforts have been made with regard to supporting the features of the IP Differentiated Services model over connection oriented networks, such as work done by the ATM Forum for ATM networks. From a customer's perspective, there is a need to have the capabilities provided by the IP Differentiated Services model in those cases where some portion of the end to end network includes an ATM network. Network service providers are faced with the need to offer multiple virtual trunks or VPNs, e.g. one virtual trunk per customer, such that each customer has access to features of the IP Differentiated Services model over such
Hunt Douglas H.
Northcote Bruce S.
Fujitsu Network Communications, Inc.
Trinh D.
Weingarten Schurgin, Gagnebin & Lebovici LLP
LandOfFree
Supporting multiple application traffic types over... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Supporting multiple application traffic types over..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Supporting multiple application traffic types over... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2965026