Reduced packet header in wireless communications network

WO 99/45678

PCI7US99/03380

REDUCED PACKET HEADER IN WIRELESS COMMUNICATIONS NETWORK

Field of the Invention

The present invention relates to communication networks and more particularly to communications networks which utilize packet communication protocols such as the Ethernet protocol.

Background of the Invention In computing and communications it has become commonplace to network computers or other data processing devices. One problem with networking processing systems relates to the infrastructure necessary to communicate between processors.

Typically, communication occurs over a hardwired interface utilizing a communication protocol such as the Ethernet protocol. However, this infrastructure is not always available or, it may be impractical in certain environments or for certain applications to create a fixed infrastructure. For example, in a residential building it may be impractical or prohibitively expensive to wire the structure for a computer network. Similarly, in a business environment, portable processing systems may need to be moved from

20-03-2000

US 009903380

<img file="AU747484B2_D0001.tif" />

-2location to location while still maintaining a link to a network.

As a result of these and other such situations, wireless and other communication systems which reduced the amount of fixed infrastructure have been utilized as the mechanism for communication between processing systems. However, wireless systems have typically been either expensive to operate or had reduced bandwidth in comparison to hardwired systems. While a dedicated wireless system may be operated at increased speed, often, these wireless networks are integrated with existing hard-wired networks and, thus, it may be desirable to use the existing networking protocols such as Ethernet, token ring or the networking protocols. However, the Ethernet protocol and other conventional networking protocols were not developed for use in a wireless environment and may be inefficient when operated in a wireless network environment.

United States Patent No. 5,446,736 describes a method and apparatus for connecting a node to a wireless network using a standard protocol. The '736 patent describes a connection oriented system where tokens are used to replace portions of a header during the connection. The article Protocols De Transmission De Donnees X 25 in Micro Systemes, no. 74, 1 April 1987, pages 94-100, provides an examination of data transmission protocols in High Level Data Link Control (HDLC). The article Low-Loss TCP/IP Header Compression

For Wireless Networks in Wireless Networks, vol. 3, no.

REPLACEMENT PAGE £

AMENDED SHEET

US 009903380

20-03-2000

<img file="AU747484B2_D0002.tif" />

<img file="AU747484B2_D0003.tif" />

-2/1-

5, 1 October 1997, pages 375-387, describes header compression for TCP/IP communications.

In light of the above discussion, a need exists for increased performance in wireless communications so as to more efficiently allow operation of existing communication protocols over wireless communication links.

Summary of the Invention

In view of the above discussion, it is an object of the present invention to provide improved efficiency in networking computers over wireless and other reduced bandwidth communication links.

REPLACEMENT PAGE

<img file="AU747484B2_D0004.tif" />

AMENDED SHEET

A further object of the present invention is to provide an increase in throughput over reduced bandwidth communication links which is transparent to users of the communication links.

Still another object of the present invention is to provide increased 5 throughput of a communication link while remaining compatible to existing networking protocols.

A further object of the invention is to alleviate at least one of the problems of the prior art.

In accordance with a first aspect of the invention there is provided a 10 method of reducing overhead in communication of messages having a message header transmitted over a communication link, where a message to be transmitted over the communication link is received and a reduced data message and an associated alias address transmitted on the communication link, the • .

’··;·. method including the steps of:

• · · · . ’··. 15 determining if an unused alias address from a first predefined set of alias :*. addresses is available;

associating an unused alias address from the first predefined set of alias • · addresses with at least a portion of the header of the received message if an ’*·.· unused alias address is available so as to provide the associated alias address;

• · associating an alias address from the first predefined set of alias addresses which is currently associated with the header of a previously received message with at least a portion of the header of the received message so as to associate a used one of the predefined set of alias addresses with the header to • · · , *.: provide the associated alias address if an unused alias address is not available;

and removing at least a portion of the message header from a subsequent message having a message header, a portion of which is substantially identical the message header of the received message, to provide the reduced data message.

<img file="AU747484B2_D0005.tif" />

3a

In accordance with a second aspect of the invention there is provided a system for reducing overhead in communication of messages having a message header over a communication link, where a message to be transmitted on the communication link is received and a reduced data message and an associated alias address transmitted on the communication link, including means for determining if an unused alias address from a first predefined set of alias addresses is available;

means for associating an unused alias address from the first predefined set of alias addresses with at least a portion of the header of the received 10 message if an unused alias address is available so as to provide the associated alias address;

means for associating a used alias address from the first predefined set of alias addresses with at least a portion of the header of the received message so ’«··. as to provide the associated alias address if an unused alias address is not • · · · ' .*’··. 15 available; and • · means for removing at least a portion of the message header from a : subsequent message having a message header, a portion of which is • · .

substantially identical the message header of the received message, to provide the reduced data message. • ·

In accordance with a third aspect of the invention there is provided a :°computer program product for reducing overhead in communication of messages Σ : having a message header over a communication link, where a message to be transmitted on the communication link is received and a reduced data message • · · ·, *·· and an associated alias address transmitted on the communication link, the computer program product including:

a computer-readable storage medium having computer-readable program code means embodied in said medium, said computer-readable program code means including:

computer-readable program code means for determining if an unused alias address from a first predefined set of alias addresses is available;

<img file="AU747484B2_D0006.tif" />

3b computer-readable program code means for associating an unused alias address from the first predefined set of alias addresses with at least a portion of the header of the received message so as to provide the associated alias address if an unused alias address is available;

computer-readable program code means for associating a used alias address from the first predefined set of alias addresses with at least a portion of the header of the received message so as to provide the associated alias address if an unused alias address is not available; and computer-readable program code means for removing at least a portion of the message header from a subsequent message having a message header, a portion of which is substantially identical the message header of the received message, to provide the reduced data message.

<img file="AU747484B2_D0007.tif" />

-4By associating an alias address with header information of a message, the alias address may be substituted for the header information. This substitution may result in fewer bytes of data being transmitted over the communication link. Furthermore, by automatically selecting used links which have not been recently used the present invention eliminates the 10 need to take down a link which has previously been established. Thus, no overhead is associated with removing an unused link.

In a particular embodiment of the present invention, the alias address and at least a portion of 15 the message header associated with the message are transmitted on the communication link. A confirmation of the association of the portion of the message header and the associated alias address may also be received from the communication link. A transmitted reduced data message may then be received from the communication link and the subsequent message reconstructed from the portion of the message header associated with the received alias address and the reduced data message.

By associating the alias address with a portion of the header of the original message, and then reconstructing the original message based on the alias address, the present invention provides for reduced overhead per communication. The reduction in overhead

<img file="AU747484B2_D0008.tif" />

may be achieved for any protocol with a header which

WO 99/45678

PCT/US99/03380

-5- .

identifies multiple packets in a message independent of the protocol of the original message because the original message is reconstructed to conform to the protocol but the format of the reduced data packet transmitted on the communication link may be independent of the format of the original message provided that the header information may be separated from the data portion of the message.

In another embodiment of the present invention, alias addresses associated with at least a portion of a message header are identified as used. However, used alias addresses may be associated with a portion of a message header by determining when each used alias address from the first predefined set of alias addresses was last utilized so as to provide a last time of use for each of the used alias addresses. A used alias address with the oldest last time of use is then associated with the portion of the header of the received message if an unused alias address is not available. Thus, the present invention may avoid the need to take down no longer used alias address associations.

In one particular embodiment of the present invention, the portion of the message header removed to create the reduced data message includes a source address and a destination address of the message.

In another embodiment of the present invention, the association of alias addresses with the message headers is monitored. The number of alias addresses in the first predefined set of alias

WO 99/45678

PCT/US99/03380

-6addresses is increased based upon the monitoring so as to provide a second predefined set of alias address. By monitoring the association of alias addresses, the present invention may selectively increase the number of available addresses to reduce the amount of communication bandwidth utilized to establish alias address associations. Thus, the number of available addresses may be dynamically adjusted to reduce the likelihood of churning of alias addresses.

In a further embodiment of the present invention, alias addresses are removed from the second predefined set of alias addresses if the alias addresses are determined to be inactive. In particular, where the second predefined set of alias addresses comprises a combination of the first predefined set of alias addresses and a third predefined set of alias addresses, the number of alias addresses may be reduced to the first predefined set of alias addresses if all of the third predefined set of alias addresses have been removed from the second predefined set of alias addresses. Thus, the number of alias addresses may also be dynamically reduced if the number of alias addresses needed to avoid churning declines.

As will further be appreciated by those of skill in the art, the present invention may be embodied as a method, apparatus/system or computer program product.

WO 99/45678

PCT/US99/03380

-7- .

Brief Description of the Drawings

Figure 1 is block diagram of a communication system utilizing one embodiment of the present invention;

Figures 2A through 2C are illustrations of data packets utilized by one embodiment of the present invention;

Figure 3 is a flowchart illustrating operations carried out upon receipt of a packet by a communications interface according to one embodiment of the present invention;

Figure 4 is a flow chart illustrating operations carried out upon receipt of data from a communication link by a communications interface according to one embodiment of the present invention;

Figures 5A through 5E are illustrations of data packets utilized by an alternative embodiment of the present invention;

Figure 6 is a flowchart illustrating operations carried out upon receipt of a packet by a communications interface according to an alternative embodiment of the present invention;

Figure 7 is a flow chart illustrating operations carried out upon receipt of data from a communication link by a communications interface according to an alternative embodiment of the present invention; and

Figure 8 is a flow chart illustrating operations carried out upon receipt of data from a communication link by a communications interface according to a second alternative embodiment of the present invention.

-8DETAILED DESCRIPTION OF THE EMBODIMENTS

The present invention now will be described more fully hereinafter with reference to the accompanying drawings, in which preferred embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and 10 will fully convey the scope of the invention to those skilled in the art. Like numbers refer to like elements throughout. As will be appreciated by one of skill in the art, the present invention may be embodied as methods or devices. Accordingly, the present invention may take the form of an entirely hardware . embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects.

Figure 1 illustrates one embodiment of the present invention. As seen, in Figure 1, a source 10 20 for a communication message has access to a wireless communication interface 12 . A destination of the communication message has access to a second wireless communication interface 14 which communicates with the first wireless communication interface 12 over wireless communication link 16.

<img file="AU747484B2_D0009.tif" />

In a particular embodiment of the present invention the source 10 and destination 18 are Ethernet sources and destinations. In such a case the communications interfaces 12 and 14 may be wireless Ethernet adapters or other adapters capable of

WO 99/45678

PCT/US99/03380

-9transmitting Ethernet packets over a wireless communication link. The source and destination devices may be network computers, workstations, personal computers, personal information managers or other devices capable of acting as the source or destination of a data packet in a packet communication protocol such as the Ethernet communication protocol.

While the present invention is described herein with reference to the Ethernet protocol, as will 10 be appreciated by those of skill in the art, other packet based communication protocols may also benefit from the teachings of the present invention. For example, Transport Control Protocol/Internet Protocol (TCP/IP), Asynchronous Transfer Mode (ATM) protocol and 15 other such packet communication based protocols may be utilized with the teachings of the present invention.

Furthermore, the present invention is described herein with reference to wireless communication links, however, the present invention 20 should not be construed as limited to wireless communications. As used herein the term communication link refers to any type of link between two processors remote from each other which allows communications between the two processors. The teachings of the present invention may be beneficial to any communication link where the amount of data to be

<img file="AU747484B2_D0010.tif" />

transmitted over the communication link is to be reduced. Examples of such communication links include low-speed modem communications, satellite communication links, cellular communication links, radio frequency

WO 99/45678

PCT/US99/03380

<img file="AU747484B2_D0011.tif" />

-ιθ- .

communication links, microwave communication links or any communication link with a rate schedule based on the amount of data transferred across the communication link.

Figure 1 illustrates a single source and a single destination which may communicate through communication link 16. However, as will be appreciated by those of skill in the art, multiple sources and destinations may be accessible through communication interfaces 12 and 14. Furthermore, the communication between a source and a destination may be in either direction across communication link 16. Communication interfaces 12 and 14 may each provide access to both sources and destinations which utilize communication interfaces 12 and 14 to communicate across communication link 16. Accordingly, Figure 1 is exemplary and should not be construed as limiting the present invention.

In overview, the operation of certain embodiments of the present invention provides for data reduction in the transmission of packets over a communication link, such as the wireless communication link 16, by reducing the header size of the communication packets. More specifically, data packets with routing information, such as destination and source addresses, are stripped of this header information prior to transmission on the communication link. An alias address is assigned to the header and the alias address is added to the data packet. As used herein the term alias address refers to a designation

WO 99/45678

PCT/US99/03380

-11- .

associated the header information of the original message such that the original message may be recreated knowing the alias address. The data and alias address are then transmitted on the communication link. The receiving side of the communication link receives the modified data packet and reconstructs the original data packet based on the alias address received. Subsequent messages with the same source and destination addresses are then also stripped of the source and destination address information and the associated alias address used for communication over the communication link.. Because the alias address can be smaller than the corresponding source and destination address of the original header, the amount of data transmitted over the communication link may be reduced.

In specific aspects of the present invention, alias addresses are assigned to unused alias addresses until all of the alias addresses have been utilized.

After all of the alias addresses have been utilized, 20 when a data packet with a new source and destination is received for transmission on the communication link, the least recently used alias address is reassigned to the new source and destination addresses. Thus, the present invention may alter the alias address associations without the overhead of explicitly terminating previous alias address associations. Furthermore, if it is determined that the assignment of alias addresses is occupying too large an amount of the bandwidth of the communication link, the size of the

<img file="AU747484B2_D0012.tif" />

WO 99/45678

PCT/US99/03380

-12alias address may be expanded to reduce the rate at which reassignment occurs.

The specific operation of various embodiments of the present invention will now be described with respect to an Ethernet example and the block diagram of Figure 1. Reference is also made to Figures 3, 4 and 6 through 8 which are flowchart illustrations of embodiments of the present invention. It will be understood that each block of the flowchart illustrations, and combinations of blocks in the flowchart illustrations, can be implemented by computer program instructions. These program instructions may be provided to a processor to produce a machine, such that the instructions which execute on the processor create means for implementing the functions specified in the flowchart block or blocks. The computer program instructions may be executed by a processor to cause a series of operational steps to be performed by the processor to produce a computer implemented process such that the instructions which execute on the processor provide steps for implementing the functions specified in the flowchart block or blocks.

Accordingly, blocks of the flowchart illustrations support combinations of means for 25 performing the specified functions, combinations of steps for performing the specified functions and program instruction means for performing the specified functions. It will also be understood that each block of the flowchart illustrations, and combinations of

<img file="AU747484B2_D0013.tif" />

blocks in the flowchart illustrations, can be

WO 99/45678

PCT7US99/03380

-13implemented by special purpose hardware-based systems which perform the specified functions or steps, or combinations of special purpose hardware and computer instructions .

Figure 3 illustrates the operation of a communications interface, such as the wireless communications interface 12, when a data packet is received for transmission on a communication link, such as the wireless communication link 16. As is illustrated in Figure 3, when the wireless communication interface 12 receives the data packet from source 10, the wireless communication interface 12 extracts the header from the data packet and determines the source and destination addresses from that header (block 30). It is then determined if an existing link (i.e. an alias address) has been associated with the source and destination address across the communication link 16 has been established for that source and destination address (block 32). If a link has not been established the least recently used alias address is assigned to the source and destination address (block 34) and a request is sent on the communication link 16 to assign the alias address to the source and destination address (block 36). The least recently used alias address will be an unused alias address if all alias addresses have not been assigned. In the present example, the alias address is one byte (8 bits) in size with 2 addresses (00 and 01) being reserved for control so as to define a predefined set of 254

<img file="AU747484B2_D0014.tif" />

alias addresses such that 254 alias addresses would

WO 99/45678

PCT/US99/03380

-14need to be assigned before a previously assigned alias address (used alias address) would be reused.

Figure 2B illustrates a data packet transmitted on the communication link 16 by a 5 communication interface 12 or 14 to request that a link be established. The format of the data packet includes 00 as the first byte of the data packet after the frame header. The 00 indicates that the packet is requesting that a link be established. The next 12 10 bytes of the packet specify the source and destination address of the original data packet. The source and destination addresses are then followed by one byte which specifies the requested alias address to be associated with the source and destination addresses.

Utilizing the format illustrated in Figure 2B a link request is a 16 byte packet.

Returning to Figure 3, after transmitting the request to establish a link, the communication interface 12 and 14 sets a reassignment flag (block 38) 20 which is used as part of a handshake with the receiving communication interface 14 or 12 to verify that the link is established (i.e. both sides of the communication link 16 have the same association of source and destination address to alias address) and waits for an acknowledgment packet to be received from

<img file="AU747484B2_D0015.tif" />

the communication link 16. After receiving the acknowledgment, the communication interface 12 or 14 checks to see if the reassignment flag remains set (block 40). If the reassignment flag remains set then the request for the link failed and a new alias address

WO 99/45678

PCT/US99/03380

-15is selected (block 34) and the process repeated (blocks 36 and 38). This process may be repeated until a link is established or may repeated for a finite number of attempts and, if still unsuccessful, the message may be 5 discarded or an error notification provided to the source of the packet.

If the reassignment flag is not set (block

40), then a link has been established. The communication interface 12 or 14 then checks the CRC of 10 the original packet (block 42) and discards the packet if a CRC error is indicated (block 44) . Alternatively, the CRC check could be performed prior to establishing a 1ink.

If no CRC error is detected, then the CRC and source and destination addresses are removed from the packet and replaced with the alias address (block 46).

A format for the packet to be transmitted on the communication link 16 in the present Ethernet example is illustrated in Figure 2A. As seen in Figure 2A, the 20 packet transmitted across the communication link 16 includes the frame header followed by the alias address (1 byte), followed by the message type (2 bytes) and then 46 to 576 bytes of data. After creation of the reduced data packet, the reduced data packet is 25 transmitted on the communication link 16 (block 48).

The processing of subsequent data packets received by a communication interface 12 or 14 having the same alias address is also illustrated in Figure 3.

<img file="AU747484B2_D0016.tif" />

As is seen in Figure 3, when the packet is received the header is extracted, (block 30) and it is determined if

WO 99/45678

PCT7US99/03380

-16a link exists (block 32). Because the packet is a subsequent packet, a link exists and the reassignment flag will not be set (block 40). The CRC of the packet is checked (block 42) and if no error is indicated the 5 header information is removed (block 46) and the reduced data packet transmitted on the communication link 16 (block 48).

Figure 4 illustrates the operations of a communication interface 12 or 14 according to one 10 embodiment of the present invention when a packet is received from the communication link 16. When a packet is received, the first byte (C) after the frame header 7E is extracted (block 50) and it is determined if the first byte after the frame header specifies a request to establish a link (C=00) (block 52). If the first byte after the frame header specifies a link request, the source, destination and alias address fields of the packet (see Figure 2B) are extracted (block 54) and the alias address is associated with the 20 source and destination addresses (block 56) . A confirmation packet is then constructed (block 58) and scheduled for transmission on the communication link 16 (block 60).

Figure 2C illustrates a data packet transmitted on the communication link 16 by a communication interface 12 or 14 to confirm that a link has been established. The format of the data packet includes 01 as the first byte of the data packet after the frame header (C=01). The 01 indicates

<img file="AU747484B2_D0017.tif" />

that the packet is confirming that a link has been

WO 99/45678

PCT/US99/03380

-1710

<img file="AU747484B2_D0018.tif" />

established. The next 12 bytes of the packet specify the source and destination address of the original data packet. The source and destination address are then followed by one byte which specifies the alias address which has been associated with the source and destination addresses. Utilizing the format illustrated in Figure 2C a link confirmation packet is a 16 byte packet.

Returning to Figure 4 at block 52, if the first byte after the frame header of the packet is not a 00 it is determined if the byte is a 01 (block 62) indicating that the packet is a confirmation packet as illustrated in Figure 2C. If the first byte after the frame header specifies a link confirmation packet then the source, destination and alias address fields of the packet (see Figure 2C) are extracted (block 64) and the reassignment flag associated with the alias address is reset (block 66). Processing then continues as described above with respect to Figure 3 to transmit the data of the original packet over the communication link 16.

If the first byte after the frame header is not 00 or 01 then the packet is a data packet of the format specified in Figure 2A. The first byte after the frame header then specifies the alias address of the packet which is matched with the source and destination address pair associated with the alias address (block 68). The original packet is then reconstructed and scheduled for transmission to the original destination 18 (block 72).

WO 99/45678

PCT/US99/03380

-18Figure 5A through 8 illustrate an alternative embodiment of the present invention where the size of the alias address may be expanded. Such an embodiment reduces the likelihood of churning of alias addresses. Because there is some overhead associated with establishing a link which associates the alias address with the source and destination addresses a condition may arise where a large percentage of the communication link 16 bandwidth is taken up by packets which establish links. The onset of this condition, referred to herein as churning, may be delayed or prevented by expanding the alias address field to include two bytes of address data, thus substantially increasing the number of available alias addresses and reducing the frequency of alias address reassignments. Figures 5A and 5B illustrate examples of the single byte and double byte addresses associated with the alternative embodiment of the present invention.

As seen in Figure 5A, the first bit of the single byte address is set to 0 to indicate the single byte address. Figure 5B illustrates that the first bit of the extended address field is set to 1 to indicate an extended address .

Figure 6 illustrates the operation of the alternate embodiment of the present invention when receiving a packet for transmission on the communication link 16. As seen in Figure 6, the communication interface 12 or 14 first determines if an existing link is associated with the source and destination addresses of the received packet (block

<img file="AU747484B2_D0019.tif" />

WO 99/45678

PCT7US99/03380

-1980). If no link exists, then it is determined if an unused alias address is available (block 82). If extended alias addresses are not in use then this determination is made based on the availability of single byte alias addresses. If extended addresses are in use then this determination is made based upon the availability of an extended alias address. If no unused alias address is available then it is determined if alias addresses are churning (block 84).

The determination of whether churning is occurring may be made by tests based on the ratio of link-establishment requests to data packets transmitted over the communication link or the time since the full address space was reassigned (for example, the time lapsed during the most recent 254 reassignment of 8-bit aliases), or other methods known to those of skill in the art. In one embodiment, churning is declared when the probability (as estimated by relative frequency of occurrence) becomes too great that an active link will be interrupted by alias reassignment, where a link is considered active if it was used during last T milliseconds .

If churning is occurring, then the field length of the alias address field is extended to more 25 than one byte and extended aliases are used (block 86) . An alias address is then selected from the unused extended alias addresses (block 90). If churning is not occurring, then the least recently used alias is selected for association with the source and destination addresses (block 88). If extended alias

<img file="AU747484B2_D0020.tif" />

WO 99/45678

PCT/US99/03380

-20addresses are not in use then this selection is made based on the last use of single byte alias addresses.

If extended addresses are in use then this selection is made based upon the last use of an extended alias address.

In either case, once an alias address is selected a link establishment request is sent on the communication link (block 92). The format of exemplary link establishment request packets are seen in Figure 10 5D for an extended alias address and Figure 2B for a single byte address. Figure 2B is described above. Figure 5D illustrates a data packet transmitted on the communication link 16 by a communication interface 12 or 14 to request that an extended alias address link be 15 established. The format of the data packet includes 02 as the first byte of the data packet after the frame header. The 02 indicates that the packet is requesting that an extended alias address link be established. The next 12 bytes of the packet specify 20 the source and destination address of the original data packet. The source and destination address are then followed by two bytes which specify the requested alias address to be associated with the source and destination addresses. Utilizing the format illustrated in Figure 5D a link request is a 17 byte packet.

After transmitting the link establishment request, the communication interface 12 or 14 waits to receive confirmation that the link has been established (block 94). The operations for establishing a link

WO 99/45678

PCT/US99/03380

-21(block 92 and 94) may be carried out as described above with reference to Figures 3 and 4 where a retry on error technique is used to establish the link.

Once the link is established the original data packet is altered as described above to remove the source and destination addresses and the CRC and to include the alias address so as to provide a reduced data packet (block 96). This reduced data packet is then sent on the link 16 (block 98). The format of exemplary reduced data packets are illustrated in Figure 2A and Figure 5C.

Figures 7 and 8 illustrate operations of alternative embodiments of the present invention which utilize extended alias addressing when a packet is received from the communication link 16. As seen in Figure 7, when a packet is received from communication link 16, the first byte after the frame header is extracted (block 100). If the first byte is a 00 or 02 then the packet is a link establishment request and the appropriate source, destination and alias addresses are extracted from the packet (block 102 and block 104) based on the format of the packet (See Figure 2B and Figure 5D). The extracted alias address is then associated with the source and destination addresses (block 106). A confirmation packet is then constructed (block 108) and scheduled for transmission on the communication link 16 (block 110).

The format of the confirmation packet corresponds to the format of the establishment request.

Accordingly, if a 00 link establishment request is

<img file="AU747484B2_D0021.tif" />

WO 99/45678

PCT7US99/03380

-22received, then the 01 format confirmation packet illustrated in Figure 2C is constructed. The confirmation packet format of Figure 2C is described above. However, if the link establishment request 5 received was a 02 extended alias address link establishment request then a confirmation packet having the format of Figure 5E is constructed.

Figure 5E illustrates a data packet transmitted on the communication link 16 by a communication interface 12 or 14 to confirm that an extended alias address link has been established. The format of the data packet includes 03 as the first byte of the data packet after the frame header. The 03 indicates that the packet is confirming that an extended alias address link has been established. The next 12 bytes of the packet specify the source and destination address of the original data packet. The source and destination address are then followed by two bytes which specify the requested alias address to be associated with the source and destination addresses. Utilizing the format illustrated in Figure 5E a link confirmation packet is a 17 byte packet.

Returning now to Figure 7, if the first byte after the frame header of the packet received from the 25 communication link 16 is not. a 00 or 02, then it is determined if the first byte after the frame header is 01 or 03 (block 112). If the first byte is a 01

V Ί \\-0 kill

Wto^' or 03 then the packet is a confirmation packet and the appropriate source, destination and alias addresses are extracted from the packet (block 112 and block 114)

WO 99/45678

PCT/US99/03380

<img file="AU747484B2_D0022.tif" />

based on the format of the packet (See Figure 2C and

Figure 5E). The reassignment flag associated with the extracted alias address is then reset to indicate that that the link was successfully established (block 116) .

If the first byte after the frame header of the received packet is not a 00, 01, 02 or 03 then the packet is a data packet of one of the formats shown in Figure 2A or Figure 5C. The alias address maybe determined from the data packet by examining the first bit of the first byte after the frame header. If that bit is a 0 then a single byte alias address is used (see Figure 5A), however, if that bit is 1 then an extended alias address is used (see Figure 5B). In either case the alias address is extracted from the packet and matched with the associated source and destination pair (block 118). The original packet is then reconstructed, including the CRC (block 12 0) and scheduled for transmission to the destination (block 122) .

Figure 8 illustrates the operation of an alternative embodiment of the present invention which utilizes extended alias addressing when a packet is received from the communication link 16 but does not utilize separate request and confirmation packets for extended alias addresses. By utilizing the same methodology to determine if churning is occurring, both sides of the communication link 16 may determine if churning is occurring independently, thereby avoiding the need for explicit signaling that extended alias addresses are to be used.

WO 99/45678

PCT/US99/03380

-24As seen in Figure 8, when a packet is received from communication link 16, the first byte after the frame header is extracted (block 130). If the first byte is a 00 or 01 the packet is either a 5 link establishment request or a confirmation packet and it is then determined if churning is occurring (blocks 132 and 142) which would indicate whether or not single byte or extended alias addressing is used. Alternatively, the receiving communication interface 10 could examine the first bit of the alias address field to determine if it was a 0 indicating single byte or a 1 indicating extended alias addressing. The communication interface could also determine the total length of the packet and, if that length is 16 bytes 15 the single byte alias addressing is used. If the packet is 17 bytes, then extended alias addressing is used. In either case, after it is determined if single byte or extended alias addressing is used, the appropriate source, destination and alias addresses are 20 extracted from the packet (block 140 or block 144) based on the format of the packet (See Figures 2B and 2C or Figure 5D and 5E) .

After extracting the alias address, it is then determined if the first byte after the frame header is 00 indicating a link request packet (block 146). If not, then the packet is a link confirmation and the reassignment flag associated with the alias address is reset (block 148). If the first byte after the frame header is a 00 then the packet is a link

<img file="AU747484B2_D0023.tif" />

establishment request and the alias address is assigned

WO 99/45678

PCT/US99/03380

-25to the source and destination (block 150) and the appropriate confirmation packet constructed (block

152). The confirmation packet is then scheduled for transmission on the communication link 16 (block 154).

Returning to block. 132, if the first byte after the frame header is not a 00 or a 01 then the packet is a data packet. The alias address is extracted as described above utilizing the first bit of the address to determine if the address is extended (block 134). The original packet is then reconstructed (block 136) and scheduled for transmission to the destination (block 138).

The association of alias addresses with source and destination addresses may be made using any 15 number of techniques for association of items. For example, a look-up table or a linked list may be used to create the association between the alias address and the source and destination addresses. In a table implementation the alias address, source and destination addresses and time of last use of the alias address may be maintained at each side of the communication link 16 so as to facilitate selection of the previously used alias address which has been used the least recently.

In the extended alias address embodiment of the present invention, a trash-collection routine periodically examines the time of last use for each link that has an extended-length alias, and long-idle links with extended addresses are removed from the tables at both ends of the communication link 16. In

<img file="AU747484B2_D0024.tif" />

WO 99/45678

PCT7US99/03380

<img file="AU747484B2_D0025.tif" />

-26- .

this way, alias length drifts back to its shortest option. Once no extended-field aliases are present in the tables, use of the 02 and 03 messages is held in abeyance until churning is once again detected, and the bridge tables are returned to their pre-expansion size .

Alternatively, the trash-collection routine could periodically remove from the tables at both ends links which have a time of last use older than a predefined threshold. When the total number of links falls below the number of links available in the set of non-extended alias addresses, the extended alias address links could be re-assigned to non-extended alias addresses for subsequent communications. As will be appreciated by those of skill in the art in light of the present disclosure, other methods of returning the system to operating with non-extended alias addresses could also be utilized.

While the present invention has been described with respect to the Ethernet communication protocol, as will be appreciated by those of skill in the art, the teachings of the present invention may be applied to other communication protocols. For example, if the communication interfaces 12 and 14 perform IP routing functions and the Ethernet packets contain IP datagrams, the communications interfaces 12 and 14 could also strip the IP routing information from within the datagrams. In such a case the routing information could either be saved or recalculated to provide the next hop in the datagrams path. Thus, the present

WO 99/45678

PCT/US99/03380

-27invention can be used with TCP/IP to further reduce overhead where the Ethernet and IP addressing are stripped and the gateway or router at each end rebundles the IP datagram in an Ethernet packet which specifies the next hop in the IP path.

In the drawings and specification, there have been disclosed typical preferred embodiments of the invention and, although specific terms are employed, they are used in a generic and descriptive sense only 10 and not for purposes of limitation, the scope of the invention being set forth in the following claims.