Multiplex communications – Pathfinding or routing – Switching a message which includes an address header
Reexamination Certificate
2001-02-21
2004-12-07
Pezzlo, John (Department: 2662)
Multiplex communications
Pathfinding or routing
Switching a message which includes an address header
C370S401000, C709S238000
Reexamination Certificate
active
06829238
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
Conventionally, the Internet was configured by wired networks typified by Ethernet. In recent years, however, networks configuring the Internet have been diversified, and an IP network using a cellular phone, a PHS, a wireless packet device, etc. has been widely used. To effectively use a given bandwidth in the Internet including wireless networks, an IP router with which an efficient transmission rate can be obtained in a connection between hosts in the case where networks having different natures such as wired and wireless are linked.
The present invention relates to an IP router device having a function for linking different networks and for terminating a TCP connection relaying IP packets between networks in a computer network based on the Internet Protocol (IP).
2. Description of the Related Art
In a wired network, an IP packet loss or delay occurs mainly due to a congestion within an IP router which links networks. In the meantime, in a wireless network, data that is modulated into radio waves in the physical layer is not properly transmitted by being influenced by a state change caused by radio wave attenuation, interference, blocking by an object, etc. so that an IP packet loss or delay occurs.
Currently, the protocol called TCP (Transmission Control Protocol) is used as a standard for making a reliable data transmission in the Internet.
However, as to TCP, it is known that a congestion occurring within a wired network is recognized as a main cause of a packet loss/delay, and an efficient transmission rate cannot be obtained in a wireless network unless the parameters and algorithms for controlling TCP are optimized for a wireless network.
Considered as a solution to the above described problem is an application gateway which uses control parameters and algorithms, which are respectively suitable for TCP connections, by once terminating TCP connections in the TCP layer and by establishing respective TCP connections for networks of different natures, and provides an efficient transmission rate.
Configuration of a conventional IP router device and that of a conventional TCP terminating device are respectively shown in
FIGS. 1 and 2
. The IP router device shown in
FIG. 1
comprises network drivers
11
a
and
11
b
, IP stacks
12
a
and
12
b
, and an IP forwarding
16
. The IP router passes IP packets between networks as shown in FIG.
1
. The header configuration of an IP packet, that of a TCP packet, and that of an ICMP packet are respectively shown in
FIGS. 3
,
4
, and
5
.
In the header of the IP packet, src-IP (SOURCE IP ADDRESS) indicating the IP address of a transmission source, and dst-IP (DESTINATION IP ADDRESS) indicating the IP address of a connection destination are described as shown in FIG.
3
. In the header of the TCP packet, src-port number (SOURCE PORT) indicating the port number of a transmission source, and dst-port number (DESTINATION PORT) indicating the port number of a connection destination are described as shown in FIG.
4
. In the header of the ICMP packet, TYPE (type) and an IP header and first 64 bits of a datagram are described.
The flow of the process for inputting an IP packet, which is performed by the IP stacks
12
a
and
12
b
of the above described conventional IP router device, is shown in
FIG. 6
, whereas the flow of the process for outputting an IP packet, which is performed by the IP stacks
12
a
and
12
b
of the IP router device, is shown in FIG.
7
. In the input process, the IP stack determines whether or not the dst-IP address of an IP packet is addressed to its local host as shown in FIG.
6
. If the IP stack determines that the IP packet is addressed to the local host, it passes the packet to a TCP/UDP stack of its local host. If the IP stack determines that the IP packet is not addressed to the local host, it passes the packet to the output process via the IP forwarding.
Additionally, in the output process, the IP stack searches a routing table with the dst-IP address or its subnetwork, and passes the IP packet to a network driver according to the routing table, as shown in FIG.
7
.
Since the IP router device only passes an IP packet between networks as described above, fundamentally, an end host is unconscious of the existence of the router. It is possible to make an IP packet uniquely reach between arbitrary hosts via this router device, and there is no change in the information of the src-IP address and the dst-IP address within the header of the IP packet. Additionally, even if a change occurs in the configuration of a router or the number of routers on an end-to-end path, there is no influence.
Namely, one of the natures of the Internet is that each host has a globally unique IP address, a communication can be made from the host to an arbitrary host, and a communication can be made from the arbitrary host to the corresponding host. This nature is called “end-to-end global connectivity guarantee”.
In the meantime, the conventional TCP terminating device (application gateway) is composed of network drivers
11
a
and
11
b
, IP stacks
12
a
and
12
b
, TCP stacks
13
a
and
13
b
, a stream transferring unit
14
, and a database
20
for holding connection information.
The flowchart of the stream transfer function of the conventional TCP terminating device (application gateway) is shown in FIG.
8
. As shown in this figure, the stream transfer function waits for the link of a TCP connection as a server socket (step S1), and establishes a TCP connection with a client (step S2). Then, the stream transfer function obtains the IP address and the port number of the connection destination from the data within a stream (step S3), and establishes a TCP connection at the connection destination as a client socket (step S4).
Then, the stream transfer function determines whether or not the stream continues (step S5). If the stream does not continue, the process is terminated. If the stream continues, the stream transfer function reads data the amount of which is equal to or smaller than a predetermined amount from the stream from the client (step S6), and writes the data to the stream to the server (step S7).
The above described TCP terminating device (application gateway) has a problem such that the global connectivity guarantee cannot be made for the following reasons.
(a) Many protocol-dependent application gateways make TCP termination only for a particular protocol. Therefore, an IP packet or a TCP stream does not flow with an unsupported application.
(b) Since a gateway that does not store end-to-end information makes TCP termination, the host of the gateway is misidentified as an end host as opposed to a partner end host. Namely, the end hosts cannot identify their counterparts mutually.
One of the important things to implement the TCP terminating device (application gateway) is that the information of a connection destination must be notified from a client host by some means or another.
As a specific example, http handling web data supports a communication via an application gateway referred to as a proxy. With http, the information of the IP address or the port number of a connection destination can be written.
However, since such a method is not supported by all of application protocols, it can be said that service scalability is significantly low.
In the meantime, there is a method using software (middleware) for a client host, which operates in coordination with an application gateway.
This software is intended to direct all TCP connections toward an application gateway, and, therefore, its service scalability is significantly improved in comparison with a normal proxy.
However, this software cannot cope with a protocol to be described below, according to which a connection is established in two stages, and a second connection is established in a reverse direction of a first connection.
Taken as a specific example is the use of a data transfer application ftp by an application gateway.
First of all, it is possible to establish a control session from a client
Chikada Michiyasu
Hisanaga Takanori
Kuwata Daisuke
Nakagawa Itaru
Taniguchi Tetsuya
Pezzlo John
Staas & Halsey , LLP
LandOfFree
IP router device having a TCP termination function and 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 IP router device having a TCP termination function and a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and IP router device having a TCP termination function and a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3313938