Multiplex communications – Fault recovery
Reexamination Certificate
1999-09-20
2003-07-15
Chin, Wellington (Department: 2664)
Multiplex communications
Fault recovery
C370S218000, C370S401000
Reexamination Certificate
active
06594227
ABSTRACT:
FIELD OF THE TECHNOLOGY
The present invention relates to a communication control system wherein communication stations which communicate with each other according to TCP/IP protocols are connected to Ethernet buses.
BACKGROUND OF THE TECHNOLOGY
TCP/IP (Transfer Control Protocol/Internet Protocol) is a globally popular protocol for computer networks. It is a combination of two protocols TCP and/or UDP and IP, where TCP is designed to accurately transfer data while IP is designed to transfer data between networks in the form of packets.
Ethernet is a LAN (Local Area Network) most frequently used for computer networks. And “Ethernet” is a registered trademark of Xerox Corporation.
In
FIG. 1
, reference symbol “B” refers to the bus. In
FIG. 2
, the structure of the communication frame “TCP MESSAGE” comprises the TCP header (HEADR) and the TCP data; the IP PACKET comprises the IP header (HEADR) and the IP data; and the ETHERNET FRAME comprises the ETHERNET header (HDR), the ETHERNET data, and the ERROR CONTROL. “TCP” means “transfer control protocol”. “IP” means “Internet Protocol”.
FIG. 1
is a conceptual diagram showing the configuration of a communication control system where a communication station which communicates with other stations according to TCP/IP protocols is connected to an Ethernet (hereinafter referred to as E bus). This diagram conceptually shows how software and hardware are configured in the communication control system.
A communication station ST is connected to an E bus B through network adapter NA.
There are TCP and UDP in the layer above the layer of IP. The communication station ST communicates with other stations according to these protocols.
UDP (User Data Protocol) is used to send a packet of data. Communication between a transmitting communication station and a receiving communication station takes place only once in an attempt to transfer such a packet, and there is no ability for this communication to recover from any communication failure. TCP establishes a connection between the transmitting communication station and the receiving communication station to provide such abilities as recovery from communication failure. Both of these protocols use an IP address and a protocol port number to recognize a mutual communication partner.
There are user application program(s) AP above the layer of TCP and UDP.
FIG. 2
is a schematic representation showing the structure of a communication frame for TCP. Network adapters NA receive only those frames directed at their local communication station in order to reduce the burden of software located just above the layer of the network adapters. For this purpose, a 6-byte receiver's address (referred to as a MAC address) is appended to the Ethernet header shown in FIG.
11
. The MAC address is specified so that it is unique to each network adapter and there is no address duplication on a global scale. The unit of data transferred using TCP/IP is called an IP packet, which is encupsulated in the data field of an Ethernet frame. TCP/IP communication is designed so that data can also be transferred using other media, such as ATM (Asynchronous Transfer) or WAN (Wide Area Network). Each IP packet, therefore, has an Internet address. (called an IP address) as the receiver's address.
As described above, if communication stations which communicate with each other according to TCP/IP protocols are connected to an E bus through network adapters, the stations are required to have both a MAC address and at least one IP address. However, IP recognizes only one MAC address for each IP, which it is communication with. This has been a bottleneck when a communication system needs to be made redundant.
Some distributed control systems are provided with an E bus and some of these distributed control systems have a redundant configuration to more than one E buses in order to increase reliability. The fact that IP recognizes only one MAC address for each remote IP is a drawback, however, when a distributed control system needs to be-connected to redundant E buses, as described below.
FIG. 3
shows examples of conventional systems with redundant connections to E buses through TCP/IP protocols. In
FIG. 3
, diagram (a) is an example where TCP is used and diagram (b) is an example where UDP is used.
Referring to
FIG. 3
, buses B
1
and B
2
are installed as E buses. In each of the transmitting and receiving communication stations, the layers ranging from network adapters NA to TCP or UDP are made dual-redundant. It is the responsibility of a user application program AP to determine which of two redundant communication routes to use. Since the IP address is accompanied with only on MAC address, the IF layer, as well as the layer of TCP or UDP just above the IP layer, must also be made dual-redundant. In this communication control system, redundancy control must be done by the user application program AP.
This example of the prior art has the following problems:
[1] Redundancy control must be carried out by the user application program, resulting in an increase in the burden of the program.
[2] Duplication of the TCP layers requires resources (memory) for connection management, for example, to be doubled.
[3] TCP is a protocol for stream-type communication, where flow control mechanism checks if the each part of stream data is securely transferred. As a result, it is relatively difficult to guarantee the complete data transfer without any loss or duplication when TCP connection is switched over.
[4] Communication for the purpose of maintaining TCP connection is essential; therefore, the redundant connections increase the burden of a network traffic.
[5] Duplication of UDP layers eliminate the problem discussed in item 4 above. However, this solution involves a problem that the order of data items becomes unpredictable.
[6] Duplication of UDP layers do not solve the problem that data, if lost, cannot be recovered.
[7] A recovery process is therefore required for the user application program to be able to solve the problem discussed in item 6 above. However, this solution significantly increases the burden of the program.
[8] Standard TCP/IP applications do not assume such redundancy as discussed above; therefore, most TCP/IP applications, except some special ones, cannot benefit from the advantages of redundant configuration.
The present invention is intended to solve the aforementioned problems. The object of the present invention, therefore, is to provide a communication control system which is configured so that only one MAC address, which IP is aware of, is passed to IP even if duplicated components are switched between, thus making it unnecessary for the user application program to be aware of any redundant configuration even if such a configuration is applied when communication stations, which communicate with each other according to TCP/IP protocols, are connected to E buses.
DESCRIPTION OF THE INVENTION
As the means for solving the aforementioned problems, the present invention provides a communication control system configured as described below.
(1) A communication control system, wherein communication stations are connected to a network through network adapters; the communication stations are provided with communication control means; the communication control means communicate with each other according to TCP/IP protocols; a specific MAC address is assigned to each of the network adapters; and the communication control means is aware of only one MAC address for any single remote communication station, having:
a configuration in which only one unit of the communication control means is incorporated in each communication station and network adapters are incorporated redundantly in each communication station, and comprising:
a redundancy control means, which exists between the network adapters and the communication control means, and which converts a receiver's MAC address included in a transmission request into the MAC address of one of
Akiyama Chuuji
Ogawa Toshi
Chin Wellington
Fox Jamal A.
Kojima Moonray
Yokogawa Electric Corporation
LandOfFree
Communication control system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Communication control system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Communication control system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3090109