Multiplex communications – Data flow congestion prevention or control – Flow control of data transmission through a network
Reexamination Certificate
1998-02-05
2002-07-16
Marcelo, Melvin (Department: 2733)
Multiplex communications
Data flow congestion prevention or control
Flow control of data transmission through a network
C370S248000, C370S395520
Reexamination Certificate
active
06421321
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an apparatus and a method for transferring a packet flow in a communication network. A conventional network for use within an enterprise or an Intranet has developed with an emphasis on transferring mails and news over a wide area, sharing resources (e.g., files and printers) and improving business efficiency within a work group.
In such a network, terminals in a group are interconnected via a common transmission path called a “local area network” (called LAN) including the “Ethernet”. A typical network system is formed by interconnecting the LANs of respective work groups through a backbone network such as Fiber Distributed Data Interface (called FDDI). The traffic in such a network system consisted mainly of so-called local traffic including data transmitted to/from the shared files and the printers.
In recent years, as the Internet is spreading to wider and wider areas and as multi-media applications are becoming more and more popular with the increasing access to the WWW (World Wide Web) service, the traffic resulting from work-group LANs accessing external resources is increasing rapidly. Further, even middleware is suffering changes as seen in the increase of applications (e.g., voice and animation) requiring high transmission speed and quality and in the development of multi-media handling protocols.
Such changes in recent years are causing performance limits to the conventional information forwarding protocols particularly in a backbone network, i.e., the “hop-by-hop best-effort” forwarding methods through routers. Accordingly, a new method and apparatus for forwarding information in the backbone network is in great demand.
2. Description of the Related Art
A new technology for transferring packets by using the address of the Network Layer Protocol (e.g., IP address of the TCP/IP protocol) in the communication network, esp. in the NBMA network (typically the ATM network) are currently being studied by international organizations such as the ATM forum (an industrial group for standardizing the ATM LAN). The conventional method for transferring packets are divided roughly into following three types.
FIGS. 1A-1C
illustrate the conventional packet transfer method in a communication network. The three methods are explained below, taking as an example in which the ATM network is used for a communication network and the TCP/IP is used for a high-layer protocol.
(1) Subnet-Relay Type
FIG. 1A
illustrates a subnet-relay type packet transfer method, where a single ATM network including three logical subnetworks (hereinafter simply called a subnetwork or subnet) is shown. The subnet relay type, which is the most popular in a so-constructed network, connects subnetworks
130
with each other through routers
110
provided therebetween. The known LAN emulation method and the classical IP-over-ATM method are of this type. Here, an end system
111
is a general term for a terminal connected directly to the ATM network and a router located at an entrance/exit (or ingress/egress) to/from the ATM network from a legacy LAN or another type of network.
An IP packet is divided into ATM cells and forwarded directly between end systems
111
which belong to the same subnetwork. However, if the end systems belong to different subnetworks (as shown in
FIG. 1A
as end systems A and B), the ATM cells are transmitted from a subnetwork to another through router
110
provided therebetween (as shown by a heavy line) even when the end systems are included in a single ATM network in which the packet is transmitted on an ATM cell basis. The ATM cells received by the router are once converted into an IP packet for relay processing and again to the ATM cells to be sent to the ATM network. That is, the relay processing is conducted in the IP layer. It is a problem that this type requires time since the relay processing is executed mostly by software.
(2) Router-cut-through Type
FIG. 1B
illustrates a router-cut-through type, which is provided with a router between subnetworks
130
as in the aforesaid subnet-relay type. However, it has an ATM cell switching function in the router. An ATM switch
113
, provided in router
112
, performs first the relay processing on the packet received through a default route in the IP layer as in the aforesaid subnet-relay type. The default route is previously provided corresponding to destinations, to pass therethrough a packet without a route specified. The default route is shown by a dotted line, which corresponds to that shown by the heavy line in the aforesaid subnet-relay type.
Since the relay processing by software requires time as mentioned above, when detecting such a packet flow as that of the FTP (File Transfer Protocol) and HTTP (Hiper Text Transfer Protocol), for which a short-cut path is useful, router
112
establishes an SVC (Switched Virtual Circuit) between the input and output ports of ATM switch
113
. The packets input to router
112
thereafter are routed through the thus-established SVC and are relayed through the heavy line by cutting through ATM switch
113
, thus speeding up the packet relaying.
(3) Cut-through-route-setting Type
FIG. 1C
illustrates a cut-through-route-setting type, which has a router provided between subnetworks
130
as in the aforesaid relay types (1) and (2). First, a packet is processed in a Next Hop Resolution Server
115
(hereinafter simply called a server), which has the same router function as in the aforesaid relay types (1) and (2), and is forwarded through the default route shown by the dotted line in FIG.
1
C.
When recognizing the packet flow as that of the aforesaid FTP and HTTP, an end system
116
(end system A), which is the ingress system for the packet flow, establishes a direct SVC between end system A and other end system
116
(end system B) which is the egress system for the packet flow so as to bypass the router (server). Succeeding IP packets are converted into cells in end system A and, forwarded in the ATM layer through the direct SVC shown by the dotted line in FIG.
1
C. However, since ingress system A is not aware of the ATM address of egress system B in this type, system A makes an ATM address resolution request by sending an NHRP Request message including the destination IP address, to a next-hop-resolution-protocol (called NHRP) server
115
(server A here) through the default route.
If server A is not aware of the ATM address of end system B, server A sends the NHRP Request message including the destination IP address, through the default route to the NHRP server
115
(server B here) which is connected to the adjacent subnetwork. When replied with the ATM address of end system B (address resolution reply) by server B, server A replies the end system A's request with the ATM address. End system A requests the ATM network to set a direct SVC between end systems A and B by using the replied ATM address.
Therefore, a mechanism to resolve the ATM address of the egress system based on the destination IP address is needed in this type.
Comparing the three relay types, types 2 and 3 are similar to type 1 in that packets are forwarded through the default route. However, the former two are superior to the latter in the relaying performance since the relaying route in the backbone-network, which passes through routers (called a hop-by-hop relay), is bypassed by the ATM connection.
Comparing types 2 and 3, they differ from each other in the method of setting the bypass route, as seen in the figures. In type 2, the default route and the bypass route are the same although packets are processed in in different ways. Type 2 has difficulties distributing the traffic to plural routes and designing network reliability, since the default route is fixed by the destination and therefore, the bypass route is fixed by the location of the router in the network. To prevent this, even if the bypass mechanism is realized with a high-speed and high-performance router, a further measure is required in the neighboring ATM network
Ogawa Jun
Sakagawa Kazuo
Fujitsu Limited
Katten Muchin Zavis & Rosenman
Marcelo Melvin
LandOfFree
Apparatus and a method for transferring a packet flow in a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and a method for transferring a packet flow in a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and a method for transferring a packet flow in a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2884176