 |
Claims  |
|
|
We claim:
1. A communication system for inter-connecting a plurality of frontend
networks with a loop-type backbone network, the communication system
comprising:
the backbone network including a plurality of logical or physical highways;
at least one node means connected to the backbone network, said node means
including a bridge means connected to a corresponding one of the frontend
networks, said bridge means including
a segmenting means for segmenting a first data block of variable length
received from the frontend network, into at least one of second data
blocks of preselected, fixed length to be sent along the backbone network,
a searching means for searching a routing information table with a
destination station address contained in said first data block,
a registering means for registering routing information including a source
station address of said first data block in said routing information
table, and
an assembling means for reassembling said first data block from at least
one of said second data blocks received from the backbone network; and
a switching means for selectively connecting said bridge means to at least
one of said backbone network highways to transmit said second data blocks
along the backbone network.
2. The communication system according to claim 1, wherein said switching
means connects said bridge means to all of said backbone network highways
in order to broadcast forward said second data blocks in accordance with
results of said searching means searching the routing information table.
3. The communication system according to claim 1, wherein each of said
second data blocks has a learning indicator field and an error check
field.
4. The communication system according to claim 1, wherein said node means
includes a plurality of said bridge means.
5. The communication system according to claim 4,
wherein said switching means connects each of said bridge means to a
predetermined one of said highways in order to receive said second data
block from the backbone network.
6. The communication system according to claim 1, wherein said bridge means
directly transmits at least one of said first data blocks reassembled by
said assembling means to the frontend network connected to said bridge
means when a transmission becomes ready.
7. The communication system according to claim 1, wherein
said assembling means includes a receive controller and a re-assemble
buffer means, said re-assemble buffer means having a re-assemble
management table and a set of fixed length buffers with a chain pointer,
and
said receive controller includes means for storing said second data blocks
consecutively received from said bridge means, and means for executing
assembling said second data blocks transmitted from the bridge means to
said first data block by using said chain pointer of said fixed length
buffers and said re-assemble management table.
8. The communication system according to claim 1, wherein each of said
second block has a backbone MAC header including at least a cell position
field and a sequence number field, and said assembling means includes a
receive controller for executing an error check of said cell position
field and sequence number field.
9. The communication system according to claim 1, wherein said assembling
means includes a receive controller having a group address register means
for storing group address and a receive decision means for deciding to
selectively receive one of said broadcasted second data blocks from said
backbone network in accordance with said group address and destination
address included in one of said broadcasted second data blocks so as to
limit a number of said bridge means which can receive one of said
broadcasted second data blocks.
10. A communication system for transferring data between a plurality of
frontend networks with a loop-type backbone network, each of the frontend
networks being connected with at least one station, the communication
system comprising:
the backbone network including a plurality of physical or logical highways;
a plurality of nodes connected to the backbone network; and,
each of the plurality of nodes including bridge means corresponding to
frontend networks and a switching means for selectively connecting said
bridge means to said highways, said bridge means including:
a first converting means for converting a first data block of variable
length received from a corresponding frontend network into at least one
second data block of fixed length, said first data block having a
destination station address and a source station address, each second data
block having error check information and learning indicator information to
be transferred to the backbone network,
a second converting means for converting at least one of said second data
blocks received from the backbone network into said first data block, and
a searching and registering means for searching a routing information table
with said destination station address to determine at least one of said
highways to be connected by said switching means on which to transmit said
second data block and for registering routing information including said
source station address in said routing information table.
11. The communication system according to claim 10, wherein said second
data block is transferred from the backbone network with said error check
information and said learning indicator information thereof, and
said searching and registering means executes a learning routine in
response to said learning indicator information and an output of an error
checking means.
12. The communication system according to claim 10, wherein each of said
second data blocks has an AFC field including at least a learning
indicator field, said learning indicator field being used as said learning
indicator information.
13. The communication system according to claim 10, wherein each of said
second data blocks has a backbone MAC header including at least a cell
position field, said cell position field being used as said learning
indicator information.
14. A method for transferring data between a plurality of frontend
networks, each of the frontend networks connecting at least one station
through bridge circuits to a backbone network, the backbone network
including a plurality of physical or logical loop highways, the method
comprising the steps of:
receiving a first data block of variable length having a destination and
source station address from one of the frontend networks;
segmenting the first data block into at least one of fixed length second
data blocks each including a control field having a learning indicator
field, a content field, and an error check code field;
determining on which highway to transmit each second data block by
searching a routing information table with said destination station
address;
transmitting each second data block on the determined highway; and
registering routing information including said source station address of
the corresponding segmented first data block in said routing information
table.
15. The data transferring method according to claim 14, further comprising
steps of:
checking said error check code field of said second data block transmitted
from the backbone network.
16. The data transferring method according to claim 15, further comprising
steps of:
learning said routing information derived from said second data block in
accordance with said learning indicator field thereof and a result of said
checking step.
17. In a communication system in which a plurality of frontend LANs which
connect a plurality of stations are connected through a plurality of
physical or logical loop highways, a node comprising:
a loop access control means for selecting the logical highways on which to
transmit a plurality of fixed length backbone LAN cells to another node;
and
a bridge means connected between said loop access control means and
corresponding frontend LANs, said bridge means including a converting
means for converting each variable length frontend LAN frame received from
one of the frontend LANs into at least one of said backbone LAN cells,
each of said backbone LAN cells having a learning indicator field and an
error code check field.
18. The node according to claim 17, wherein said learning indicator field
has information denoting whether said cell is a first, single one of said
backbone LAN cells.
19. The node according to claim 18, further comprising:
a storing/searching means for storing routing information including a set
of a station address and a node address, said storing/searching means
storing said routing information in response to said learning indicator
field when said node receives said backbone LAN cells from the physical or
logical loop highways.
20. The node according to claim 13, wherein a plurality of said bridge
means are connected to said loop access control means, and said routing
information further includes a corresponding bridge address.
21. A communication system for communicating among a plurality of stations
on a plurality of frontend LANs via a loop-type backbone LAN, the
communication system comprising:
the backbone LAN including a plurality of physical or logical highways;
a bridge means connected to each of the frontend LANs, said bridge means
including:
a segmenting means for segmenting a variable length first block from the
frontend LAN to at least one of second data blocks, each of said second
data blocks having a fixed length, and
a searching means for searching a routing information table with a
destination table; and
a switching means for selectively connecting said bridge means to at least
one of said logical highways to transfer said second data block to another
one of the bridge means.
22. The communication system according to claim 21, wherein said bridge
means further includes a re-assembling means for re-assembling said second
data blocks received from the backbone LAN into said first data block.
23. A communication system for communicating among a plurality of stations
on a plurality of frontend LANs via a backbone LAN, the communication
system comprising:
a backbone LAN including a plurality of physical or logical highways;
a bridge means connected to each of the frontend LANs, said bridge means
including:
a segmented means for segmenting a variable length first data block from
the frontend LAN to at least one of second data blocks, each of said
second data blocks having a fixed length, and
a searching means for searching a routing information table with a
destination station address contained in said first data block; and
switching means for connecting said bridge means to all of said physical or
logical highways to transfer said second data blocks by broadcast in
accordance with a searching result of said searching means.
24. In a communication system between a plurality of stations connected to
a plurality of frontend networks, each of the frontend networks being
connected to one of a plurality nodes of a backbone network that includes
a plurality of logical or physical loop highways, the one node having a
routing information table, a method of learning routing information in the
node comprising the steps of:
segmenting a frontend network frame received from the frontend network into
at least one backbone network cell to be transmitted to the backbone
network;
searching the routing information table to retrieve a search result
corresponding to a destination station address contained in said frontend
network frame; and
selecting one or all of the highways in response to the result of said
searching step on which to transmit said backbone network cell to the
backbone network.
25. In a communication system between a plurality of stations connected to
a plurality of frontend networks, each of the frontend networks being
connected to one of a plurality of nodes of a backbone network that
includes a plurality of logical or physical loop highways, in which each
node has a table for storing routing information, a means to convert first
data block received from the frontend network to at least one second data
block, and a means to transmit each second data block to a designated one
of the nodes, a method of learning the routing information in the node
comprising the steps of:
searching the table with a destination station address contained in each
first data block received from the frontend network;
adding learning indicator information to each second data block, which
learning indicator information includes station position information;
transmitting each second data block with said added learning indicator
information to a corresponding one of the highways;
receiving each second data block with said added learning indicator
information from all of the highways; and
registering said station position information from the learning indicator
information of each received second data block into the table.
26. A communication system for interconnecting a plurality of frontend
networks with a loop-type backbone network, the communication system
comprising:
the backbone network including a plurality of logical or physical highways;
at least one node means connected to the backbone network, said node means
including a bridge means connected to a corresponding one of the frontend
networks, said bridge means includes
a segment means for segmenting a first data block of variable length
received from the frontend network, into second data blocks of fixed
length to be transmitted along the backbone network, each of said second
data blocks having a destination address field, a source address field, a
backbone MAC header field, and a cell content field,
a searching means for searching a routing information table with a
destination station address contained in said first data blocks, and
an assembling means for re-assembling said first data block from at least
one of said second data blocks received from the backbone network, said
assembling means including a receive control means and a reassemble buffer
means for buffering reassembled said first data block, and said receive
control means having means for storing a group broadcast information and
means for selectively receiving broadcasted second data blocks in
accordance with said destination field and said group broadcast
information; and,
a switching means for selectively connecting said bridge means to at least
one of said backbone network highways to transmit said second data block
along the backbone network. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
BACKGROUND OF THE INVENTION
The present invention relates to a bridge apparatus connecting between LANs
(Local Area Networks) and a communication system between networks using
the apparatus, and more specifically to a bridge apparatus connecting
between a LAN transferring in fixed data unit and a control method
thereof.
In a backbone LAN incorporating a frontend LAN, a routing system and a
control method thereof in the prior art is disclosed in Japanese patent
application laid-open No. 196132/1988 and performs as follows.
In communication between stations connected to frontend LANs, each station
need not know the routing information regarding frontend LAN-ID to which a
destination station is connected, nor its node address of a backbone LAN.
A sending station address and a destination station address are added to
the transmission data. If the destination address is in the same LAN as
that of the sending station, the destination station can receive this data
directly. Also if the destination station is connected through the
backbone LAN to another frontend LAN, communication can be performed with
transmittivity as if the transmitting and receiving stations were in the
same LAN. This routing control is performed basically in accordance with
the algorithm of learning bridge specified in the IEEE 802.1d MAC (Media
Access Control) Bridges by nodes constituting LANs as follows.
A node holds an entry table where positions of stations are registered.
When the node performs retrieval from the entry table regarding a
destination station address of the frame supplied by a frontend LAN, and
if the destination station is in the frontend LAN connected to the self
node, the received frame is discarded. If the destination station is not
in the frontend LAN connected to the self node or if it cannot be found,
the received frame is transmitted by broadcast to the backbone LAN. Each
backbone LAN node receives this frame and performs retrieval of the
destination address from the entry table. If the destination station
address is in the frontend LAN incorporated by the self node, the received
frame is forwarded to the frontend LAN incorporated by the self node
allowing the destination station to receives this frame. If the entry
cannot be found, the frame is forwarded to the frontend LAN. If the
destination station is in a frontend LAN incorporated by another node, the
received frame is discarded.
As above described, the entry table, deciding whether a frame is forwarded
using a station address as a key, has information indicating whether the
station having the address is at the frontend LAN side or at the backbone
LAN side. The entry table possessed by each node can learn whether a
station is at the frontend LAN side or at the backbone LAN side from the
sending (source) station address of frames flowing through the frontend
LAN or frames flowing through the backbone LAN.
In the above-mentioned prior art, however, forward processing between
frontend LANs through a backbone LAN constituted by a plurality of
physical or logical links is not considered. Data transfer from the
backbone LAN to the frontend LAN cannot be performed.
Also in the prior art, forwarding between a plurality of frontend LANs
incorporated in a node is not considered and therefore data transfer
cannot be performed to these frontend LANs.
In the prior art, since learning of the routing information in a backbone
LAN in fixed length data unit is not considered, following point becomes a
problem.
That is, data received from the frontend LAN by the node of the backbone
LAN is divided in a plurality of fixed length data units. However, only
one of the data units of the backbone LAN include the routing information,
and other nodes of the backbone LAN does not recognize data to be learned.
Also in the prior art, since learning the routing information of a backbone
LAN constituted by a plurality of physical or logical links is not
considered, the following point becomes a problem.
When the node receiving the frame of the frontend LAN performs transmission
by broadcast to the backbone LAN, the same data is transmitted along all
links. On the other hand, as the learning manner in other node, the
learning must be performed from all links. In this case, the learning by
the same information occurs in the number of the links, and the futile
learning process increases.
Also in the prior art, since learning of the routing information of a
network constituted by a loop-shaped link is not considered, following
point becomes a problem.
That is, the node forwarding data again receives the data after taking a
round of the loop. Consequently, the learning is performed twice, i.e.,
when the data is received from the frontend LAN and when the data is
received from the backbone LAN, and the futile learning process increases.
In addition, the prior art relating to a bridge circuit is disclosed, for
example, in U.S. Pat. No. 4,597,078, "Bridge circuit for interconnecting
networks".
SUMMARY OF THE INVENTION
An object of the invention is to provide a bridge apparatus for connecting
a network performing communication in fixed length data unit to other
network and a control method thereof.
Another object of the invention is to provide a bridge apparatus for
connecting a network constituted by a plurality of links to other plural
networks and a control method thereof.
A further object of the invention is to provide a bridge apparatus for
connecting a network constituted by a loop-shaped link to other network
and a control method thereof.
Still another object of the invention is to provide a bridge apparatus for
connecting a plurality of frontend networks connecting stations to a
backbone network and a communication system between networks using the
bridge apparatus.
In order to attain the foregoing objects, in the invention, in constitution
that frontend LANs connecting a plurality of stations are connected
respectively to a plurality of nodes of a backbone LAN constituted by a
plurality of physical or logical links. Each node corresponds to a
frontend LAN. A first data (long data) block received from the frontend
LAN is divided and converted into one or plural second data (short data)
block units with fixed length and then transferred to a destination node.
A bridge apparatus is provided where the transferred second data blocks
are combined and converted into the first data block. The bridge apparatus
can transmit the second data blocks to arbitrary links, and the second
data blocks are received from one link. The bridge apparatus also has
means for deciding whether the learning of routing information should be
performed or not, based on the learning indication information existing in
the second data blocks including the routing (station position)
information.
That is, in the invention, a bridge apparatus is provided where a first
data block received from a frontend LAN is divided and converted into one
or more second data blocks having fixed length and error check code and
then transferred to a destination node. The bridge apparatus has means for
supplying the information indicating whether the learning is necessary or
not (learning indication information) to the second data block including
the routing information among the second data blocks, and means for
learning the routing information based on the information indicating
whether the learning is necessary or not when the second data blocks are
received.
That is, the basic principle of the invention is in that the data transfer
can be performed between a plurality of frontend LANs through a backbone
LAN constituted by a plurality of loop shaped physical or logical links.
The data is segmented to short data units with fixed lengths a.
In the invention, in a backbone LAN system comprising a plurality of
loop-shaped physical or logical links and a node device connected thereto
for transferring data to second data units with fixed length, each node
has a plurality of ports corresponding to frontend LANs and each port has
the bridge function to perform the data transfer between a plurality of
physical or logical links of the backbone LAN and the frontend LANs.
The node at the transmission side supplies the learning indication to the
short data including the routing information, and the node or port at the
receiving side performs the learning of routing information from the
learning indication of the short data.
On receiving the first data from the frontend LAN, the port registers set
sending station address of data and the self node address go and the self
port address to the entry of the entry table, and uses the destination
address of the data as a key retrieve the entry. Thereby the destination
node/port address can be obtained and the decision of discard/forwarding
of the first data is performed. The decision results are of the following
three sorts.
(1) The destination station is incorporated in the frontend LAN of the self
node/port. Consequently, the data is discarded.
(2) The destination station is incorporated in the frontend LAN of another
node. Consequently, the destination node/port address and the sending
(self) node/port address are added to all short data, and further the
short data including the sending station address is indicated by the
learning indication and transmitted to the link to which the receiving
means of the destination node/port of the backbone LAN is connected.
The port and the receiving link are preferably connected fixedly. That is,
the receiving port address and the receiving link have the same number,
thereby at the transmission side the data is transmitted to the link
having the same number as that of the destination port address, and at the
receiving side the receiving may be performed only from the link to which
the self port is connected.
(3) The decision station cannot be found in the entry; therefore state of
unknown destination occurs. Consequently, the destination node/port
address set to the broadcast address and the sending (self) node/port
address are added to all short data, and further the short data including
the sending station address is indicated by the learning indication and
transmitted to all links of the backbone LAN.
On the other hand, the receiving port receives the short data from the link
assigned to the self port, and compares the destination node/port address
with the self node/port address/ If both are coincident (including
broadcast), the short data is taken and assembled to the data for the
frontend LAN and then transferred to the frontend LAN of the self node. In
this case, the entry need not be retrieved using the destination station
address as a key.
The learning is performed when the learning indication is taken from the
short data, thereby the sending station information (sending node/port,
sending station address) is extracted and registered to the entry.
The port, after transmission of the short data with the learning
indication, is not learned here, because the short data performing one
round of the loop has been previously learned at the receiving state from
the frontend LAN. That is, the learning is not performed from the short
data where the destination node/port address and the self node/port
address are coincident.
The invention may be learned as follows.
The node takes the short data with the learning indication bundled in all
links, and transmits the sending station information to all ports in the
self node. Each port registers the sending station information to the
entry possessed by the port respectively.
At the transmission side (transmission port), the learning indication is
added to the short data indicating the sending station position. The short
data with the learning indication is then transmitted only to one link.
Even in broadcast, only one arbitrary link is sufficient for the short
data including the sending station address requiring the learning
indication (for example, the link corresponding to the self port).
Also the invention may be learned as follows. Even in the case of discard,
only the short data including the position information of the sending
station may be transferred to the frontend LAN. Thereby, all ports can
register the station position to the entry in similar manner to the case
of forwarding. After this state, each port is enabled in the decision of
forwarding/discard regarding this entry. Thereby transmission by broadcast
due to the unknown destination is decreased.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block constitution diagram illustrating an embodiment of the
invention;
FIGS. 2A-2F are diagrams illustrating an embodiment in frame constitution
of a frontend LAN and a cell constitution of a backbone LAN;
FIG. 3 is a diagram illustrating an embodiment of an entry table held by a
node (port) connected to a backbone LAN;
FIGS. 4A-4C are a block diagram of an embodiment of a switch means 1003 in
the node of FIG. 1, a block diagram of an embodiment of a learning cell
multiplexer 10034, and a cell constitution diagram illustrating operation
of the learning cell multiplexer 10034, respectively;
FIG. 5 is a block diagram of an embodiment of a transmission controller
1008 in the node of FIG. 1;
FIG. 6 is a block diagram of an embodiment of a receive controller 1006 in
the node of FIG. 1;
FIG. 7 is a block diagram of an embodiment of an FDB controller 1016 in the
node of FIG. 1;
FIGS. 8A, 8B are diagrams illustrating system constitution in an embodiment
and an application example of the invention;
FIGS. 9A, 9B are flow charts of program executed by a microprocessor 1013
in FIG. 1;
FIGS. 10A-10C are explanation diagrams of re-assemble processing executed
by a receive controller 1006 in FIG. 1 and FIG. 6;
FIG. 11 is a block diagram illustrating operation of a cell header
generator 10086 in FIG. 5;
FIGS. 12A-12D are system model diagrams illustrating another embodiment of
the invention;
FIGS. 13A-13C are diagrams illustrating frame constitution of a frontend
LAN, frame constitution of a backbone LAN and constitution of a learning
frame for the routing learning in another embodiment of the invention;
FIG. 14 is a diagram illustrating a specific example of a node connected to
the backbone LAN of FIG. 12; and
FIGS. 15A to 15D are system constitution diagrams illustrating a further
embodiment of the invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
Embodiments of the invention will now be described referring to the
accompanying drawings.
1. Constitution
1. 1 Whole Constitution
FIG. 8A is a constitution diagram of the whole system representing an
embodiment of the invention. A backbone LAN 0 comprises a loop-shaped
physical link 13 and a plurality of nodes 10 connected thereto. Frontend
LANs 25 at the outside of the backbone LAN 0 are connected through the
nodes 10 to the backbone LAN 0. To each frontend LAN 25 are generally
connected a plurality of stations (also called terminals) 251. In this
case, the frontend LAN 25 is, for example, an FDDI (Fiber Distributed Data
Interface), and the station 251 has MAC addresses of 48 bits. The frontend
LAN 25 may be LANs in the MAC system being different from each other. A
management device 211 may use a general work station having a file device.
A node with the management device 211 connected thereto is called a master
node, and a node other than this is called a slave node. The management
device 211 is connected to the master node through a LAN, for example, a
Ethernet 212. The management device 211 has the operation command
function, that is, the function of changing the configuration of the
backbone LAN in accordance with the command inputted by the operator and
collecting the statistical information within the node. It also has the
status monitor function, that is, the function of supervising the
operation state of the backbone LAN and generating the alarm for the
operator upon detecting trouble and also performing the logging for the
file device.
1. 2 Node Constitution
FIG. 1 is a block diagram illustrating constitution of an embodiment of a
node 10. In FIG. 1, an Ethernet interface 213 exists only at the master
node. The node 10 comprises a loop access controller 100 and a plurality
of ports 10A. The loop access controller 100 is connected to each port 10A
through a set of lines M, RC, R, TC, and T. Furthermore, the loop access
controller 100 is connected to all the ports 10A through the line L
commonly. The relevant invention regarding the node constitution assigned
to the same assignee as that of the present application is disclosed in Y.
Takiyasu et al., "High-speed ring LAN system", U.S. Ser. No. 07/399,901,
filed Aug. 29, 1989.
1. 2. 1 Loop Access Controller
The loop access controller 100 comprises an opto/electric converter 1001, a
demultiplexer 1002 for demultiplexing the input data into, for example,
logical highways i, j, k, l (hereinafter referred to simply as "highways")
of 155 Mbps.times.N (N:4 for example), switch means 1003 for performing
receive, transmit or forwarding of a fixed length bucket (called cell), a
multiplexer 1004 for multiplexing N highways, and an electro/optic
converter 1005. The demultiplexer 1002 and the multiplexer 1004 perform
multiplexing/demultiplexing in N (for example, N=4) of SONET (Synchronous
Optical Network) frames in so-called CCITT standard. In the demultiplexer
1002, an SOH (Selection Overhead) and a VC-4 (Vertical Container) are
demultiplexed from the SONET frames, and the cell boundary signal and data
of the cell are transmitted onto the highways. Further, numeral 1000
designates a microprocessor. Numeral 100A designates a memory such as RAM,
ROM or the like. The microprocessor 1000 can access the memory 100A and
other members (1001-1005) within the loop access controller 100 and also
the Ethernet interface 213 if any. Further the control information is
transmitted or received from the microprocessor 1000 through the switch
means 1003, thereby communication is possible between the microprocessors
1000 with different nodes 10.
1. 2. 2 Port
The port 10A has the function of "Learning Bridge" specified in the IEEE
802.1d MAC Bridges.
The port 10A comprises a receive controller 1006 where cells received from
the switch means 1003 are assembled into a frame of the frontend LANs, a
re-assemble buffer 1007, a transmit controller 1008, a cell buffer 1009
where frames of the frontend LANs are divided into cells and transmitted
in cell unit to the switch means 1003, an FDB (Filtering Data Base)
controller 1016, and a frontend LAN controller 1012. Further, numeral 1013
designates a microprocessor, and numeral 1014 designates a memory such as
RAM, ROM or the like. Further, numeral 1015 designates a node I/F thereby
the microprocessor 1013 of the port 10A and the microprocessor 1000 of the
loop access controller 100 can communicate.
The frontend LAN controller 1012 further comprises an FDDI access means
10123, a receive buffer controller 10122, serial interfaces 10124 and
10125, and a receive buffer 10121. In this case, the FDDI access means
10123 can be realized, for example, by Am79C83 (FORMAC) of AMD (Advanced
Micro Devices Inc.), and the receive buffer controller 10122 by Am79C82
(DPC) and Am79C81 (RBC) of AMD, and the serial interfaces 10124 and 10125
by Am7984 (ENDEC) and Am7985 (EDS) of AMD and optical module DM74-742-XF
of Sumitomo Electric Industries, Ltd., and the receive buffer by RAM. Also
a frame strip means 10126 is a circuit for stripping a frontend LAN frame
950 (FIG. 2A) from the FDDI ring when the frontend LAN frame 950,
forwarded (transmitted) from the backbone LAN side to the frontend LAN
side by the port 10A, runs around the FDDI ring and returns to the port
10A. The frame strip means 10126 can be realized by a stripping circuit by
forwarding frame counter control based on "Forwarding frame stripping
system" described in ANSI materials PROPOSAL ON FRAME STRIPPING FOR
BRIDGES IN FDDI (Henry Yang, K. K. Ramakrishman and Bill Hawe, Jun. 16,
1989) for example, or by a CAM (Content Addressable Memory).
2. Explanation of Backbone LAN and Frontend LAN
2. 1 Outline of Structure of Cell and Frontend LAN Frame
On each highway of the backbone LAN 0, a plurality of cells 963, as shown
in FIG. 2B, take a round. The cell 963 comprises a header, a cell content
field 961 and an ICS (Information Check Sequence) 962. The header further
comprises an ACF (Access Control Field) 955, a destination node/port
address 956, a source node/port address 957, an HCS (Header check
Sequence) 958, and a backbone MAC header 959.
On the other hand, FIG. 2A shows frame structure of frontend LAN. A
frontend LAN frame 950 comprises, for example, an FC (Frame Control) 964,
a destination station address 951, a source station address 952, a
frontend LAN information field 953 holding terminal information, and an
FCS 954.
2. 2 Details of Frontend LAN Frame
Table 1 shows an example of frame structure of frontend LAN.
TABLE 1
______________________________________
Frame Structure of Frontend LAN
Field Length Explanation
______________________________________
FC 8 bits frame sort
(SMT, LLC etc.)
destination station
48 bits in accordance with
address format of IEEE 802 MAC
address
source station
48 bits in accordance with
address format of IEEE 802 MAC
address
frontend LAN
variable information of transmit/
information field
(0 to 4500 receive between
bytes) terminals
FCS 32 bits error check code
covering from FC to
frontend LAN information
field
______________________________________
2. 3 Details of backbone LAN cell
FIGS. 2C, 2D show details of the ACF 955 and the backbone MAC header 959
respectively. Also Table 2 shows an structure example of the destination
node/port address 956, the source node/port address 957, the HCS 958, and
the ICS 962.
TABLE 2
______________________________________
Cell Structure (partly)
Field Length Explanation
______________________________________
destination node
10 bits assign address of
address destination node and port
destination port
2 bits Most significant bit (I/G
address bit) of destination node
address indicates whether
it is individual address
(I/G = 0) or broadcast
address (I/G = 1).
Both together are called
destination node/port
address
source node 10 bits assign address of source
address node and port
Both together are called
source port 2 bits source node/port address.
address
HCS 8 bits error check code
covering destination node/
port address and source
node/port address
ICS 16 bits error check code
covering backbone MAC
header and cell content
field
cell content field
44 bytes entering data to be
transferred
______________________________________
Although the node address and the port address are separated in the
embodiment in Table 2, it is easily analogized that the node address and
the port address as a whole can be treated as one node address. In this
case, in FIG. 1, the port 10A within the node 10 becomes one in number.
2. 4 Segmenting/Re-assembling to/from cells
Transfer within the backbone LAN by the cell 963 is performed as shown in
FIG. 2F, excluding the FCS 954 among the frontend LAN frame 950 (the
legnth being made Lf bytes). Also in another embodiment, transfer
including the FCS 954 is performed. At any rate, the transferring portion
within the backbone LAN by the cell 963 is called transfer frontend LAN
frame. The amount of Lf is different, but the following description can be
similarly applied. Depending on the amount of Lf, in following conditions,
segment to cell (generation of cell) is performed, and a part or the whole
of the frontend LAN frame is copied to the cell content field. In this
case, the length of the cell content field 961 is made to Lc bytes. It is
usually made to several tens bytes.
(1) If Lf.ltoreq.Lc, a single cell is generated.
(2) If Lc<Lf, cells are generated in the order of First cell at first, Next
cell of 0 or more in number at next step, a and Last cell.
Regarding only the Single cell and the First cell, the forwarding function
header 960 exists at the tip end portion of the cell content field 961.
FIG. 2B (1)-(4) shows the form of the First/Next/Last/Single cells, and
FIG. 2E shows form of the forward function header. As clearly understood
from the above description, the destination station address and the source
station address of the frontend LAN frame are stored in the cell content
field of the First cell or the Single cell. The forward function header
960 will be described later.
3. Initialization by Software
The microprocessor 1000 of the loop access controller 100 starts the
operation by reset of the power ON or the like, and by the program control
on the memory 100A comprising ROM and RAM. The initialization is performed
for the O/E means, 1001 the demultiplexer 1002, the switch means 1003, the
multiplexer 1004, the E/O means 1005, the Ethernet interface means 213 and
the like. Thereby, the node can communicate with other node through the
physical link 13.
The port 10A performs the initialization by the microprocessor 1013, the
memory 1014 for the frontend LAN controller 1012, the node i/f 1015, the
receive controller 1006, the re-assemble buffer 1007, the FDB means 1016,
the transmit controller 1008, the cell buffer 1007. Thereby, the data
transfer can be performed betwe | | |
