Patent Application
20100158008
The present application claims priority from Japanese applications JP2008-326792 filed on Dec. 24, 2008, the content of which is hereby incorporated by reference into this application.
The present invention relates to a technology for providing improved reliability and enhanced efficiency when TDMoIP packet loss occurs in an IP network in a situation where a network system and a switching exchange are used to handle both an ISDN and an IP network.
TDM (Time Division Multiplexing) over IP (hereinafter referred to as TDMoIP) achieves PRI (Primary Rate Interface) transmission, T1 transmission, and E1 transmission over an IP-based network or Ethernet (registered trademark)-based network.
A TDMoIP transmission technology mainly involves the use of two different methods. One method is to replace a TDM network and an end-user device entirely with a communication device provided with mechanisms for voice transmission and signaling. The other method is to tunnel TDM data through a packet network by using an existing PBX (Private Branch eXchange) and multiplexer as they are. Business enterprises expect existing devices to make use of an IP network with a view toward taking advantage of cost reduction by integrating their own voice and data networks without squandering the investment in a conventional PBX or TDM device.
In an apparatus that can use an existing PBX and multiplexer as they are, TDM clock jitter and wander are strictly defined for a TDM-based device. Therefore, synchronism is maintained with an extremely small delay. Meanwhile, when an IP-based packet method is used, packet delay and packet loss occur due to a conflict with bandwidth and router ports. A source device transmits packets at regular intervals over a network. However, the network does not guarantee that the packets will be delivered to a destination device at the same intervals or in the same sequence. In some cases, the packets will not be delivered to the destination device.
An ISDN (Integrated Services Digital Network) is no exception. It is anticipated that packet loss may also occur due to the connection of the aforementioned IP network in a situation where the IP network is integrated with an existing ISDN. When packet loss occurs, for example, the quality of communication between telephone terminals deteriorates. Measures for packet loss prevention during a packet transfer period of an IP network have been studied (as described, for instance, in JP-A-2007-60345) because they need to be taken when an ISDN, which is a TDM service, is to be integrated into an IP-based network.
The technology described in JP-A-2007-60345 packetizes a data stream in a network where a data stream is transmitted and received, copies the packets, and transmits the packets in multiple sessions. If a receiving end has a copied packet (having the same sequence number as another packet), which is copied in accordance with an RTP (Real Time Protocol) sequence number in a packet, such a copied packet is discarded. If only a copied packet has arrived at the receiving end due to packet loss, such a packet is decoded into a data stream.
The technology described in JP-A-2007-60345 makes multiple copies of packets which are data identical with each other. Thus, the copied packets are also equal in a destination address and other items of information within an IP header. Therefore, these packets are routed along the same path. The packets can be retransmitted by predefining the number of transmissions and the transmission intervals. Further, it is possible to prevent packet loss due to bursty high traffic. However, if the path becomes physically defective, packet loss may occur because the packets cannot be recovered. In addition, an IP header is attached to an RTP packet, which is small in data size. Therefore, particularly when multiple copies are made, it is anticipated that poor bandwidth efficiency may result because the header size is relatively larger than the data size. There is a trade-off between bandwidth efficiency and packet recovery technology.
To solve the above-described conventional problem and recover TDMoIP packets, the degree of multiplicity is varied with the delay in an IP network and the amount of TDM data to be transmitted. This increases the transfer efficiency and the bandwidth efficiency of low-delay transfer in consideration of bandwidth conditions. Further, generation copies obtained by multiplexing TDM data are used instead of mere packet copies. Therefore, IP headers are not duplicated so that TDMoIP packets are unfailingly delivered to a receiving-end switching exchange even if a trouble occurs in the same path. Consequently, it is possible to provide a technology for recovering TDMoIP packets even when packet loss occurs in an IP network.
According to one aspect of the present invention, there is provided a network system including: a transmitter for transmitting packet data, which is obtained by packetizing time-division-multiplexed data, to a network; and a receiver for receiving the packet data. The transmitter includes a first memory section for storing segment data, which is obtained by segmenting the time-division-multiplexed data; a control section for controlling the generation of duplicate data from the segment data; a first packet processing section for multiplexing the segment data and converting the multiplexed segment data to packet data; and a transmission section for transmitting the packet data to the network. The receiver includes a reception section for receiving the packet data and a second packet processing section for converting the received packet data to time-division-multiplexed data.
According to another aspect of the present invention, there is provided a communication method for using a transmitter for transmitting packet data, which is obtained by packetizing time-division-multiplexed data, to a network; a receiver for receiving the packet data; and a network for connecting the transmitter and the receiver, the communication method including the steps of: segmenting the time-division-multiplexed data received by the transmitter; storing the resulting segment data in a first memory section; reading the segment data from the first memory section and generating duplicate data; causing a first packet processing section to multiplex the segment data and convert the multiplexed segment data to packet data; transmitting the packet data from the transmitter to the network; and causing a second packet processing section to convert the packet data, which is received by the receiver through the network, to time-division-multiplexed data.
When the TDM data is to be converted to IP packets, a transmitting-end switching exchange may store the data in a TDMoIP packet payload section with the data size of a data queue section. Further, the size of the TDM data may be rendered variable as desired by data queue control provided by the data queue section. Furthermore, when a TDMoIP packet processing section converts the TDM data to IP packets, the TDM data may be multiplexed and stored in the TDMoIP packet payload section. Moreover, when the TDM data is to be multiplexed, the degree of multiplicity may be varied with the type of TDM data and the delay time of the IP network.
A protocol identification section may be rendered capable of identifying the type of TDM data. Further, a data management table in a memory may be rendered capable of calculating the delay time of the IP network from the timestamp of a TDMoIP packet and the time of TDMoIP reception by a receiving-end switching exchange.
A delay information packet generation section may transmit a delay information packet indicative of the delay time to the transmitting-end switching exchange and use the delay information packet to vary the degree of multiplicity. Further, the receiving-end switching exchange may reassemble TDM data from TDMoIP packets by causing the TDMoIP packet processing section to compensate for lost packets when packet loss occurs and discard redundant TDMoIP payload data when no packet loss occurs. Moreover, the TDMoIP packet processing section may discard redundant TDMoIP payload data when a duplicate sequence number is revealed by TDMoIP header information.
When TDMoIP packet loss occurs in an IP network in a situation where a network system and a switching exchange are used to handle both an ISDN and an IP network, the present invention makes it possible to provide enhanced reliability in packet recovery by making generation copies of TDM data and increase the transfer efficiency and the bandwidth efficiency of low-delay transfer in consideration of the delay in the IP network and the type of TDM data.
FIGS. 9(1) to 9(5) are diagrams illustrating an operation performed by a transmitting-end switching exchange according to a first embodiment of the present invention.
FIGS. 11(1) to 11(5) are diagrams illustrating an operation performed by a receiving-end switching exchange according to the first embodiment of the present invention.
FIGS. 14(1) to 14(4) are diagrams illustrating a delay information packet and an operation performed by a transmitting-end switching exchange according to a second embodiment of the present invention.
FIGS. 17(1) to 17(3) are diagrams illustrating an operation performed by the transmitting-end switching exchange according to the second embodiment of the present invention.
Embodiments of the present invention will be described below in detail with reference to the accompanying drawings.
A first embodiment of the present invention will now be described.
Various data are handled by the telephone terminal 10 and PC terminal 20. For example, the telephone terminal 10 handles continuous voice data and the PC terminal 20 handles single-shot data for email transmission/reception. All such data are passed through the TA 30 and converted to IP packets in the transmitting-end switching exchange 40. These IP packets are handled as TDMoIP packets 1000, passed through an IP network 50, converted to TDM data in the receiving-end switching exchange 41, and forwarded to the telephone terminal 11 and PC terminal 21 through the TA 31. These steps are performed to provide a conversation between the telephone terminal 10 and telephone terminal 11 and email transmission/reception between the PC terminal 20 and PC terminal 21.
When voice data is handled, the degree of multiplicity, which will be described later, is decreased in consideration of delay. When, on the other hand, email data or other common data is handled, the degree of multiplicity is increased to enhance the transfer efficiency and the bandwidth efficiency of low-delay transfer.
A TDM packet processing section 408 (418) converts the TDM data to IP packets. The TDMoIP packets 1000 are transmitted to the IP network 50 through a packet transmission configuration (a packet SW section 409 (419) and a port section 404 (414)). All the above steps are common to the transmitting-end switching exchange 40 and receiving-end switching exchange 41.
The following description applies to the transmitting-end switching exchange 40 only. The data to be converted to TDMoIP packets is temporarily queued in a data queue 40B in a memory section 40A. Data queues can be set up as desired. The number of data queues that are set represents the number of generations of copies. A data queue control section 40C controls the generation copies (duplicates) in the data queues.
Next, the TDM data needs to be multiplexed before being converted to TDMoIP packets. The degree of multiplicity is determined in accordance with the type of data as described earlier or determined in accordance with delay in the IP network as described later. Multiplexing is provided by the TDM packet processing section 408. The degree of multiplicity is determined on the basis of a data management table 40D in the memory section 40A. Multiplexed TDM data is stored in a TDMoIP payload section 1001 which will be described later with reference to
The following description applies to the receiving-end switching exchange 41 only. TDMoIP packets 1000 received through a reception configuration (the port section 414 and the packet SW section 419 in a protocol processing section 415) are disassembled in the TDM packet processing section 418 and reassembled into TDM data. Packet disassembly is achieved by removing the IP header 1005, UDP header 1004, timestamp 1003, and sequence number 1002. When the packets are disassembled, a sequence number comparison table 41E is searched for previously received TDMoIP packets 1000. When any existing sequence number coincides with a sequence number 41E-1 in the table, duplicate data in copied TDMoIP payload 10001 is discarded. The duplicate data is memorized in the field of payload internal data 41E-2, which will be described later with reference to
Delay information for determining the degree of multiplicity will now be described. A delay information packet processing section 417 transmits a delay information packet 2000 from the receiving-end switching exchange 41 to the transmitting-end switching exchange 40. The delay information packet 2000 contains delay information indicative of delay time between the transmitting-end switching exchange 40 and receiving-end switching exchange 41, which is derived from a data management table 41D in the receiving-end switching exchange 41.
The transmitting-end switching exchange 40 receives the delay information packet 2000, causes a delay information packet processing section 407 to disassemble the packet and acquire the delay information 2000-1, and writes the acquired delay information 2000-1 into the data management table 40D. The delay information is then used to calculate the degree of multiplicity, which varies with the delay in the IP network.
An operation of the present embodiment will now be described in detail.
Referring to FIG. 9(1), it is assumed that the data continuously transmitted from the telephone terminal 10 and PC terminal 20 through the TA 30 is A+B+C, and that the size of the data queue 40B is set to 3 (step 10001 of
Next, the protocol identification section 406 identifies the protocol type (step 10002 of
Referring to FIG. 9(2), TDMoIP payload is formed. As the size of the data queue 40B is 3, the queue size is 3. Therefore, up to three generations of TDM data are stored in the payload 1000-1 within a TDMoIP packet as detailed later with reference to FIG. 9(5). Two pieces of data A multiplexed into data (1), two pieces of data B multiplexed into data (2), and two pieces of data C multiplexed into data (3) are stored as individual payload data 1000-1-1 (step 10008 of
Referring to FIG. 9(3), TDMoIP packetization is performed. In addition, a copy is made at the time of data transmission. The data queue 20A has a First-In-First-Out (FIFO) structure, stores data (2) in the first queue position, and sequentially stores data (3) and data (4).
A data length value 1000-1-2, a data count 1000-1-3, a sequence number (step 10009 of
Referring to FIG. 9(4), a copy is made at the time of transmission in the same manner as indicated in FIG. 9(3). The data queue 40B stores data (3) in the first queue position, and sequentially stores data (4) and data (5). Data (1)′ and data (2) are handled as single payload data. A data length value 1000-1-2, a data count 1000-1-3, a sequence number (step 10009 of
Referring to FIG. 9(5), N (3 in FIG. 9(5)) copies are made of queued data. The data queue 40B stores data (4) in the first queue position, and sequentially stores data (5) and data (6). Data (1)″, data (2)′, and data (3) are handled as single payload data. A data length value 1000-1-2, a data count 1000-1-3, a sequence number (step 10009 of
Referring to FIG. 11(1), the receiving-end switching exchange 41 receives a TDMoIP packet 1000 (step 11001 of
Referring to FIG. 11(2), the order in which TDMoIP packets 1000 arrive at the receiving-end switching exchange is not guaranteed because the IP network 50 lies between the switching exchanges 40, 41. It is therefore assumed that the third TDMoIP packet 1000 transmitted from the transmitting-end switching exchange 40 is queued so as to store data (1)″, data (2)′, and data (3) in the second queue position (step 11003 of
Referring to FIG. 11(3), any duplicate received packet is discarded in accordance with a sequence number. As is the case described with reference to FIG. 11(1), the data queue 41B stores data (1) in the first queue position, data (1)″, data (2)′, and data (3) in the second queue position, and data (1)′ and data (2) in the third queue position (step 11003 of
Referring to FIG. 11(4), the data queue 41B stores data (2)′ in the first queue position because data (1) is deleted, stores data (2) in the second queue position, and stores data (2)″, data (3)′, and data (4) in the third queue position.
Referring to FIG. 11(5), it is assumed that TDMoIP packets 1000 are sequentially received, and that data (2) to (3) have the same sequence number. In such an instance, the data queue 41B stores data (4)″, data (5)′, and data (6) in the first queue position, stores data (5)″, data (6)′, and data (7) in the second queue position, and stores data (6)″, data (7)′, and data (8) in the third queue position. As all the data required for data (1) to (3) are readied, they are reassembled into data (A+B+C) for one frame of TDM data (step 11006 of
A second embodiment of the present invention will now be described.
Referring to FIG. 14(1), the delay time of the IP network 50, which is calculated from the data management table 41D shown in
Referring to FIG. 14(2), the degree of multiplicity is set to 3. As the degree of multiplicity is 3 and the size of the data queue 40B is 3 (N=3), up to three generations of packets are stored in the TDMoIP packet internal payload 1000-1. Three pieces of data A multiplexed into data (1), three pieces of data B multiplexed into data (2), and three pieces of data C multiplexed into data (3) are stored as individual payload data (step 10008 of
Referring to FIG. 14(3), TDMoIP packetization is performed. In addition, a copy is made at the time of data transmission. The data queue 40B stores data (2) in the first queue position, and sequentially stores data (3) and data (4). A data length value 1000-1-2, a data count 1000-1-3, a sequence number (step 10009 of
Referring to FIG. 14(4), N copies are made of queued data. The data queue 40B stores data (4) in the first queue position, and sequentially stores data (5) and data (6). Data (1)″, data (2)′, and data (3) are handled as single payload data. A data length value 1000-1-2, a data count 1000-1-3, a sequence number (step 10009 of
Referring to FIG. 17(1), it is assumed that a TDMoIP packet 1000 containing data (1) is already received by the receiving-end switching exchange 41. The reception time of the TDMoIP packet 1000 is stored in the data management table to calculate the delay time (step 11002 of
Referring to FIG. 17(2), a TDMoIP packet 1000 containing data (2)″, data (3)′, and data (4) and a TDMoIP packet 1000 containing data (3)″, data (4)′, and data (5) are delivered to the data queue because the TDMoIP packet 1000 containing data (1)′ and data (2) and the TDMoIP packet 1000 containing data (1)″, data (2)′, and data (3) are lost. In the same manner as described with reference to FIG. 17(1), the TDMoIP packets 1000 are disassembled to compare the sequence numbers (step 11004 of
The sequence number 1002, on the other hand, is duplicated. Therefore, data (3)″ and data (4)′, which are the copies of data (3) and data (4) and placed in the third queue position of the payload internal data 41E-2, are discarded (step 11005 of
The exemplary configuration according to the first or second embodiment described above makes it possible to provide enhanced reliability in packet recovery by making generation copies of TDM data and increase the transfer efficiency and the bandwidth efficiency of low-delay transfer in consideration of the delay in an IP network and the type of TDM data when TDMoIP packet loss occurs in the IP network 50 in a situation where a network system and a switching exchange are used to handle both an ISDN and IP network.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2008-326792 | Dec 2008 | JP | national |
