Distributing ATM cells to output ports based upon...

Details Distributing ATM cells to output ports based upon... Distributing ATM cells to output ports based upon...

1999-08-24

2004-08-24

Nguyen, Chau (Department: 2663)

Multiplex communications

Pathfinding or routing

Switching a message which includes an address header

C370S395520, C370S401000

Reexamination Certificate

active

06781994

ABSTRACT:

TECHNICAL FIELD
The present invention relates to an ATM relay device which constructs an IP relay function in a packet communication scheme represented by TCP/IP or the like using destination information on the basis of ATM (Asynchronous Transfer Mode), and a network including it.
BACKGROUND ART
With growth of the Internet, the Internet infrastructure keeps on developing. An important technique in this respect is a device for connecting a plurality of logical networks. A packet relay device such as a bridge, router, or the like forwards packets by routing in units of packet destinations in the data link layer in case of communications within a given subnet (logical network) or in the network layer in case of communications between different subnets (logical networks).
Packet relay using ATM uses a packet forwarding scheme called hop-by-hop forwarding.
FIG. 1
is a diagram showing the arrangement of such conventional ATM relay device, which comprises an ATM switch
160
which includes an ATM switch core
160
a
having a plurality of ports (#
1
to #N), and an IP forwarding unit
161
which has a cell/data unit segmentation and reassembly unit (SAR: Segmentation And Reassembly)
161
a
, and an IP processor
161
b
including a CPU.
In hop-by-hop forwarding, extraction of IP header information and determination of the next destination (next HOP) of a packet depend on software processes by the CPU. Hence, the processing speed of such processes is low, and it can be a bottleneck in a network of a plurality of relay devices considering recent trend for higher-speed, heavier communication traffic, thus disturbing an increase in throughput.
In recent years, a method of processing hop-by-hop forwarding of packets by hardware has been proposed. More specifically, this method achieves high-speed search by using table lookup which makes the destination field of an IP packet correspond to output line information, and distributes packets. By executing such hardware process, a higher-speed relay process than the conventional software process can be implemented.
Furthermore, in a relay device that uses an ATM switch, a cell switch router (to be referred to as a CSR hereinafter) scheme has been proposed. In this scheme, for example, upon relaying successive IP packet data, processing of the network layer in the layer
3
level is switched to high-speed forwarding using the layer
2
level called cut-through forwarding.
That is, a conventional IP router controls routing of packet data
372
using only the network layer of a plurality of layers
371
of the communication protocol, as shown in FIG.
2
. However, the CSR is a technique for forwarding packet data at high speed using the data link layer in addition to the network layer, i.e., a high-speed, high-performance router which does not require any packet forwarding by means of a CPU (software), as shown in FIG.
3
.
The CSR scheme is effective for improving the throughput since it uses cut-through forwarding (forwarding in the layer
2
level (ATM layer)) in case of communications with a relatively long session time using protocols such as ftp, http, and the like, and can greatly increase the packet forwarding speed compared to conventional routers.
VPI/VCI used in an ATM relay device will be briefly explained below.
Some virtual paths (VPs) are set in an ATM network, and some virtual channels (VCs) for actually sending data are set in units of VPs. For this reason, in order to identify a VP to be used, a VPI (Virtual Path Identifier) is assigned to the header field of a cell. Furthermore, in order to identify VCs in the VP, which are to be used for sending data to the partner, a VCI (Virtual Channel Identifier) is assigned to the header field of the cell. Since the VPI/VCI are the fields for determining VPs and VCs to be selected upon communicating with the partner, they are called routing bits (bits for selecting a communication route). Note that each cell consists of 53 bytes, 5 bytes of which are used as a header, and the remaining 48 bytes of which are used as user information.
FIGS. 4A and 4B
are views for explaining two cell forwarding schemes of the CSR.
Referring to
FIG. 4A
, upon hop-by-hop forwarding, an IP packet coming from a neighboring relay device
383
is sent to an IP processing unit
381
A via an ATM switch
381
B of a CSR relay device
381
, undergoes a software process, and is then forwarded to the next neighboring relay device
383
′. At this time, a virtual connection VC-
1
is established between the neighboring relay device
383
and CSR relay device
381
, and a virtual connection VC-
2
is established between the CSR relay device
381
and neighboring relay device
383
′. Reference numeral
382
denotes a communication protocol called an FANP normally used in the CSR, which protocol is used to establish/cancel a path for directly forwarding a packet in the data link layer (ATM-SW).
That is, when packet forwarding by the IP processing unit
381
A is switched to cut-through forwarding by means of the ATM switch
381
B by establishing an FTP session or the like, an ATM path is established to establish a cut-through path for switching to an ATM layer process, as shown in FIG.
4
B. At this time, forwarding is done by the ATM switch
381
B, and the IP processing unit
381
A does not forward any packets. Reference numeral
388
denotes a virtual connection VC-
3
established between the CSR relay device
381
and neighboring relay device
383
′ upon cut-through forwarding.
The two schemes upon cut-through forwarding in the CSR will be briefly explained below. One scheme is called topology driven, which establishes an ATM cut-through path addressed to an identical subnet, i.e., between end-to-end CSRS, and forwards a packet with a large data size at high speed for a long period of time. Otherwise, hop-by-hop forwarding is made. The other scheme is flow driven, that performs high-speed forwarding between neighboring CSRs (link-to-link CSRS). The aforementioned flow-driven scheme is used to implement cut-through processes in units of communications (flows), and has features: it can flexibly cope with changes in network configuration, and is suitable for a route that requires flexible quality control in units of flows, and a route with light packet traffic.
On the other hand, the topology-driven scheme is used to simultaneously implement cut-through processes for each partner, and has features: it is suitable for the backbone of a large-scale network, and a route which is expected to have heavy packet traffic.
The CSR scheme comprises an IP processing unit after, e.g., an ATM switch, and the IP processing unit controls cut-through forwarding by the ATM switch.
FIG. 5
shows the arrangement of an ATM relay device of the CSR scheme, which comprises an IP processing unit
290
, and an ATM switch
299
which is connected to this IP processing unit
290
via a single physical path. A main control unit
291
simultaneously processes by software (or hardware) information such as IP forwarding processes (IP packets, routing information), signaling process signal, and other control signals via a data/cell segmentation and reassembly unit (SAR: Segmentation And Reassembly)
395
inserted in this physical path.
In cut-through forwarding in the CSR scheme, high-speed signaling processes such as setup of a cut-through path with respect to a default path (a communication path in normal hop-by-hop forwarding) is also required. As for this cut-through path, a method that uses a permanent communication path like PVC (Permanent Virtual Connection) is available.
The signal processes are preferably done on demand using SVC (Switched Virtual connection) in consideration of channel resources.
FIG. 6
is a diagram showing another example of the arrangement of the conventional CSR relay device, in which an IP processing unit
331
is placed after an ATM switch
332
via a link
334
with IP processing. Upon executing ATM UNI (User Network Interface) signaling processes, a signaling cell is supplied to a data/cell segme

Affiliated with

Also associated with

Rating

Distributing ATM cells to output ports based upon... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Distributing ATM cells to output ports based upon..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Distributing ATM cells to output ports based upon... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3344503