WO2001078279A1

WO2001078279A1 - Network interface

Info

Publication number: WO2001078279A1
Application number: PCT/US2001/004276
Authority: WO
Inventors: Harsh Kapoor; Paul Gallo; Douglass Walker; Brian Myrick
Original assignee: Appian Communications, Inc.
Priority date: 2000-04-10
Filing date: 2001-02-09
Publication date: 2001-10-18
Also published as: AU2001236850A1

Abstract

A network interface for sharing a communication channel implements a method for enabling a customer to insert, into a payload containing a first data packet addressed to a first destination, a second data packet addressed to a second destination. The method includes determining the extent of available space in the first data packet and, on the basis of the extent of that space, identifying a second data packet for insertion into the payload. The network interface also implements a method for enabling a customer to retrieve a data packet from a payload when that payload includes data packets addressed to different users. In a circuit-switching network such as a SONET ring, this eliminates wasted space on the payload.

Description

NETWORK INTERFACE

This invention relates to communication networks, and in particular, to the sharing of limited network resources among a community of customers.

BACKGROUND

In a communication network, it is often useful to permit two or more customers to share a single transmission line, path, or trunk. A known method of accomplishing this is to allot a set of time intervals to a customer and to interleave those intervals among similar time intervals assigned to other customers. This technique, referred to as "time-division multiplexing," assures that no one customer monopolizes the transmission line for an extended period.

Time division multiplexing of signals is particularly common on the telephone network. In such a network, an analog signal from a customer's telephone travels to a codec where it is sampled at an appropriate sampling frequency. The resulting sampled values are then encoded as binary numbers in a process referred to as "pulse-code modulation." These samples (now encoded as binary numbers) are then placed into time intervals that are interleaved with time intervals containing samples taken from analog signals of other telephone customers. As a result, each time interval includes header information indicative of the destination of the samples contained therein. Each time interval therefore contains only one destination.

When the time intervals assigned to a particular customer have a common destination, it is useful to visualize the set of time intervals as forming a channel between a source customer and a destination customer. A telephone transmission line can therefore be visualized as having many such channels. Each such channel is dedicated to the transmission of a signal from one source customer to one destination customer.

Since its inception, the telephone network has been optimized for carrying speech. By its nature, speech results in signals with only short periods of silence. Hence, when the sampled analog signal is a speech signal, the time intervals dedicated to a particular channel are usually filled with samples. This results in the efficient use of available bandwidth on the transmission line.

Unlike speech signals, data signals are typically bursty, with long periods of silence between transmissions. It comes as no surprise therefore that a network optimized for carrying the steady flow of speech may be less efficient at carrying sporadic bursts of data. Because data traffic is bursty, time intervals dedicated to a channel carrying data are often not completely filled. In fact, in many cases, the majority of these time intervals are completely empty. It would be useful, therefore, to allocate this unused time to form an additional communication channel between another source customer and destination customer.

A difficulty associated with having another source customer use free space in a channel already dedicated to carrying information between an assigned source customer and destination customer is that the telephone network is a circuit switched network. This means that when one customer makes a telephone call to another customer, the telephone network dedicates a channel, or circuit, for the exclusive use of those parties. This channel, once established, is reserved for the exclusive use of those two customers, whether or not any data traffic exists in that channel.

SUMMARY

The method of the invention determines whether an inbound payload that already contains a data packet addressed to a first destination has space available for transmission of additional data. If such space is available, a network interface places a second, independently addressable data packet in that space. This results in a payload having two data packets that can be addressed to two different destinations. A network interface according to the invention therefore commandeers an unused portion of a channel that is already being used to send data to a first destination and uses that unused portion of the channel to send data to a second destination that may be different from the first destination.

The invention provides a method for inserting, into a payload that is already in use for carrying a first data packet addressed to a first destination, a second data packet addressed to a second destination. The method includes determining the available capacity of the payload and selecting, in response to the extent of this available capacity, the second data packet. This second data packet is then inserted into the payload for transmission on the communication network. In order to alert other network interfaces to the presence of this second data packet, the network interface alters the header associated with the data packet to conform to the existence of the additional data packet.

A network interface also provides for the recovery of the second data packet at the second destination. In this aspect of the invention, the method includes selectively removing a data packet from a payload on the basis of the first destination. This permits a network interface to remove selected data packets from the payload and to deliver those packets to customers served by that network interface. Typically, the header associated with the data packet is also altered to free space formerly occupied by that selectively removed data packet. These and other features of the invention will be-appreciated upon examining the following detailed description and its accompanying figures in which:

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 shows a SONET ring in which network interfaces according to the invention are deployed;

FIG. 2 shows a representative network interface from FIG. 1 ; and

FIG. 3 shows a representative method implemented by the network interface of FIG. 1.

DETAILED DESCRIPTION

Referring to FIG. 1, a network interface 10 incorporating the principles of the invention operates in connection with a local area network (LAN) interface 12 and a wide area network (WAN) interface 14. The LAN interface 12 is in communication with a plurality of customers 16a-d on a packet switched network. The WAN interface 14 is in communication with a trunk 18 of a wide area network. Additional network interfaces 20, 22, 24, each of which is likewise in communication with a plurality of customers 26, 28, 30, are also connected to the trunk 18. An example of a packet switch network suitable for connection to the LAN interface 12 is an ethernet. A network suitable for connection with the WAN interface 14 is a telecommunication network such as a SONET (Synchronous Optical Network) ring.

A conventional network interface examines a data envelope arriving from an upstream direction to determine if the payload of that data envelope contains data packets addressed to any of the customers served by the network interface. If it does, the network interface extracts the data. If it does not, the network interface allows the data to proceed in a downstream direction to the next network interface.

A network interface 10 according to the invention carries this further by determining whether there exists space available for transmission of data originating at that network interface. If such space exists, the network interface places new data in that time interval. In effect, the network interface 10 takes an unused portion of a channel dedicated to sending data from a source customer to a first destination customer and uses that unused portion of the channel to send data to a second destination customer. The network interface does so by providing a processing section 32 in communication with the WAN interface 14 and the LAN interface 12. The processing section 32, shown in more detail in FIG. 2, includes a time-slot processor 34 that communicates with the WAN interface 14. At each time interval, the time-slot processor 34 accepts an inbound data envelope 36 received by the WAN interface 14 during that time interval.

The data envelope 36 includes a payload 38 containing an inbound data packet 40 whose addressee is identified in a header section 42 of the data envelope 36. The inbound data packet 40 can be addressed to a local customer 16a on the local area network served by the network interface 10. Alternatively, the inbound data packet 40 can be addressed to a remote customer on a local area network served by a downstream network interface. The former is referred to as a "local inbound packet" and the latter as a "remote inbound packet." Because data communication tends to be sporadic, the payload 38 can also include an unused section 44 of sufficient extent to accommodate one or more additional data packets.

The time-slot processor 34 extracts the payload 38 from the data envelope 36 and provides it to a payload processor 46. The payload processor 46 then determines if any of the inbound packets in the payload 38 are local inbound packets. If any of the data packets are local inbound packets, the payload processor 46 provides those data packets to a packet processor 47 that includes a packet add drop unit 48. The packet add/drop unit 48 then passes those data packets to the LAN interface 12 for placement into a queue and eventual transmission to the addressee.

Concurrently, the payload processor 46 determines the extent of the unused section 44, namely the extent of the payload less any remote inbound packets that will be retransmitted downstream on the network. The payload processor 46 then communicates this information to the packet processor 47. In response, the packet processor 47 interrogates the LAN interface 12 to determine if there are any outbound data packets that are queued for transmission on the trunk 18 and if any of those outbound packets are small enough to fit into the unused section 44.

If no outbound packets satisfying the foregoing criteria are available, the packet processor 47 provides this information to the payload processor 46. In response, the payload processor 46 generates an outbound payload 50 and provides it to the time-slot processor 34. The outbound payload 50 includes all the remote inbound packets but none of the local inbound packets. Since the contents of the payload have changed, the time-slot processor 34 also generates an outbound header section 52 to conform with the contents of the outbound payload 50. The time-slot processor 34 then generates an outbound data envelope 54 having the outbound header section 52 and the outbound payload 50 and passes this outbound data envelope 54 to the WAN interface 14 for transmission on the trunk 18.

If outbound packets satisfying the foregoing criteria are available, the packet add/drop unit 48 provides those packets to the payload processor 46. The payload processor then incorporates these outbound packets into an outbound payload 50. The outbound payload 50 thus includes all remote inbound packets and as many outbound packets as will fit in the unused section 44. Since the contents of the payload have changed, the time-slot processor 34 generates an outbound header section 52 to conform with the contents of the outbound payload 50. The time-slot processor 34 then generates an outbound data envelope 54 having the outbound header section 52 and the outbound payload 50 and passes this outbound data envelope 54 to the WAN interface 14 for transmission on the trunk 18.

FIG. 3 summarizes the steps followed by the processing section 32 in implementing the method of the invention. At each cycle, the time-slot processor retrieves a data envelope (step 56) and provides it to the payload processor (step 58). The payload processor then determines if there are any inbound packets (step 60) and if there are any outbound packets (step 62). If the payload processor determines that there are the payload contains inbound packets, it enters a first loop 64 to classify those inbound packets into local inbound packets and remote inbound packets.

To implement the first loop 64, the payload processor initializes a counter (step 66) and increments it by one (step 68). The payload processor then extracts an inbound packet identified by the current value of the counter (step 70) and determines whether that inbound packet is intended for a local customer (step 72). If the packet is intended for a local customer, the payload processor passes the packet to the packet add/drop module to be dropped from the payload and queued for delivery to that local customer (step 74). Otherwise, the payload processor proceeds to the next inbound packet, if any (step 76).

If the payload processor determines that there exist outbound packets queued for transmission, it initializes a counter (step 78) and enters the body of a second loop 80.

Within the body of the second loop 80, the payload processor increments the counter (step 82) and inspects the outbound data packet identified by the current value of the counter (step 84). The payload processor then determines if the space available in the payload is sufficient to accommodate that data packet (step 86). If the space available is sufficient, the payload processor adds that outbound packet to the payload (step 88) and determines whether to re-execute the body of the second loop 80 by inspecting the queue for additional outbound packets (step 90). If the data packet identified by the counter is too large to fit into the available space in the payload, the payload processor proceeds directly to inspecting the queue for additional outbound packets (step 90).

If there are no additional outbound packets queued for transmission, the payload processor releases the payload to the time-slot manager for continued transmission on the network (step 92) and awaits the arrival of the next data envelope (step 94).

While the method and system of the invention have been described in connection with a time-division multiplexed system, the principles of the invention are equally applicable to a wavelength division multiplexed system. The scope of the invention as set forth in the following claims is independent of whether the shared network resource is an interval of time on a transmission line or an interval of frequency or wavelength on the transmission line.

Having described the invention and a preferred embodiment thereof, what we claim as new and secured by letters patent is:

Claims

1. A method for sharing a payload being used to send a first data packet to a first destination in a communication network, said method comprising :

determining an extent of an available capacity in said payload;

selecting, on the basis of said extent, a second data packet intended for a second destination; and

inserting said second data packet into said payload.

2. The method of claim 1 further comprising selecting said commumcation network to be a time-division multiplexed network.

3. The method of claim 2 wherein selecting said communication network to be a time- division multiplexed network comprises selecting said network to be a SONET ring.

4. The method of claim 1 further comprising selecting said communication network to be a wave-division multiplexed network.

5. The method of claim 1 wherein determining an extent of said available capacity comprises : retrieving a header associated with said payload; and

determining said extent on the basis of information contained within said header.

6 The method of claim 5 wherein determining an extent of said available capacity further comprises determining an extent of said first data packet.

7. The method of claim 1 wherein selecting said second data packet comprises polling a LAN interface to determine whether there exist second data packets queued for transmission.

8. The method of claim 7 wherein selecting said second data packet further comprises determining an extent of a data packet queued for transmission.

9. The method of claim 7 further comprising selecting said LAN interface to be an interface to a packet-switched network.

10. The method of claim 9 further comprising selecting said LAN interface to be an interface to an ethernet.

11. The method of claim 1 wherein inserting said second data packet into said payload comprises associating a second header with said payload to indicate the presence of said second data packet in said payload.

12. The method of claim 1 further comprising selecting said first destination to be an address of a local customer.

13. The method of claim 1 further comprising selecting said second destination to be an address of a remote customer.

14. A method for sharing a payload carrying a first data packet addressed to a first destination in a communication network and a second data packet addressed to a second destination in said communication network, said method comprising :

determining said first destination and said second destination;

selecting a data packet from said first and second data packets on the basis of said first destination and second destination;

retrieving said selected data packet from said payload.

15. The method of claim 14 further comprising selecting said communication network to be a time-division multiplexed network.

16 The method of claim 15 wherein selecting said communication network to be a time-division multiplexed network comprises selecting said communication network to be a SONET ring.

17 The method of claim 14 further comprising selecting said communication network to be a wavelength division multiplexed network.

18. The method of claim 14 wherein determining said first destination and said second destination comprises inspecting a header associated with said payload, said header including information indicative of the content of said payload.

19. The method of claim 14 wherein retrieving said selected data packet comprises removing said selected data packet from said payload.

20. The method of claim 19 wherein removing said selected packet from said payload comprises modifying a header associated with said payload to remove information indicative of the existence of said selected data packet in said payload.

21. A network interface for sharing a payload being used to send a first data packet to a first destination in a communication network, said network interface comprising:

a time-slot processor in communication with said communication network for receiving said payload from said communication network;

a payload processor in communication with said time-slot processor for determining an extent of an available capacity in said payload;

a packet processor for selecting, on the basis of said extent, a second data packet for insertion into said available capacity of said payload, said second data packet being intended for a second destination.

22. The network interface of claim 21 wherein said communication network is a time- division multiplexed network.

23. The network interface of claim 22 wherein said time-division multiplexed network comprises a SONET ring.

24. The network interface of claim 21 wherein said communication network is a wave- division multiplexed network.

25. The network interface of claim 21 further comprising a LAN interface in communication with said packet processor.

26. The network interface of claim 25 wherein said LAN interface is an interface to a packet-switched network.

27. The network interface of claim 26 wherein said LAN interface is an interface to an ethernet.

28. The network interface of claim 21 wherein said first destination is an address of a local customer.

29. The network interface of claim 21 wherein said second destination is an address of a remote customer.

30. A network interface for sharing a payload carrying a first data packet addressed to a first destination in a communication network and a second data packet addressed to a second destination in said communication network, said network interface comprising a time-slot processor in communication with said communication network for receiving said payload from said communication network; and

a payload processor in commumcation with said time-slot processor for determining said first destination and said second destination and selecting said data packet addressed to said first destination.

31. The network interface of claim 30 wherein said communication network is a time- division multiplexed network.

32. The network interface of claim 31 wherein said time-division multiplexed network comprises a SONET ring.

33. The network interface of claim 30 wherein said communication network to be a wavelength division multiplexed network.