DESC:FIELD OF INVENTION
The present invention generally relates to a system for fragmenting jumbo Ethernet packets. More specifically, the present invention relates to a system for fragmenting jumbo Ethernet packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links including Dense Wavelength Division Multiplexing (DWDM) transmission system with reduced local memory usage and reduced processing transmission latency. Further, the present invention relates to a method for fragmenting jumbo Ethernet frames or packets to send over TDM/OTN links.

BACKGROUND OF INVENTION
Fragmentation is a phenomenon in which storage space is used inefficiently, reducing capacity or performance and often both. The exact consequences of fragmentation depend on the specific system of storage allocation in use and the particular form of fragmentation. Jumbo frames are Ethernet frames with more than 1500 bytes of payload. Conventionally, jumbo frames can carry up to 9000 bytes of payload, but variations exist. Many Gigabit Ethernet switches and Gigabit Ethernet network interface cards support jumbo frames. Some fast Ethernet switches and fast Ethernet network interface cards also support jumbo frames.

Conventionally, Ethernet frames from one or more channels are received by Ethernet receiver engine (EthRxE) and copied into memory. Using suitable method, these frames or packets are read from memory and sent over Optical Transport Network (OTN) link after doing Generic Framing Procedure (GFP) or any appropriate framing by GFP Transmitter engine (GFPTxE). GFP framed packets received at the receiver are copied into memory of each channel. The Ethernet transmitter engine (EthTxE) of each channel then transmits over Ethernet link. The GFP receiver engine (GFPRxE) receives the entire packet and copies into specific channel EthTxE memory. This method requires full packet memory at EthTxE and Ethernet receiver engine (EthRxE) and the packet latency is as high as two packet time over Ethernet. When Ethernet arrival rate and overhead is more than the available bandwidth from the OTN, full packet handling at TX and Rx is required.

Some of the prior arts are:
US8811428 discloses a method for passing ingress and egress jumbo frames and maximizing the use of frame bursts in a cable modem which comprise of determining whether the burst frame has available space, when space is available, fragmenting a packet of the one or more packets into at least a first and a second fragment packet, and adding the first fragment packet to the burst frame and transmitting the burst frame during a transmit window. However, fragmenting is done after determining whether the burst frame has available space. Further, the fragmented size cannot be chosen based on the excess bandwidth available. The cited document does not disclose about controlling errors on data links.

EP1489782 discloses a data transfer method for reducing data transfer delays in intermediate network devices in a wide area Ethernet network, where a Media Access Control (MAC) frame having a length of longer than or equal to N bytes is divided into divided frames each having a length of M bytes (N > M) in edge devices which form a wide area Ethernet network. However, the drawback is size of the fragmented frames cannot be chosen based on the excess bandwidth available although the data transfer method has reduced data transfer delays in intermediate network devices in a wide area Ethernet network.

US7343540 discloses a method for sending data packets from a transmitter to a receiver through a shared-medium digital transmission network, each packet ending in a stop symbols field delimiting the end of said each packet, the method comprising the steps of: dividing data of at least a portion of said each packet into one or more frames; applying a systematic forward-error-correction (FEC) block code to each frame of said each packet, said systematic FEC block code maintaining data symbols, of the frames, in an unaltered state and adding parity-check symbols. However, the fragmented size cannot be chosen based on the excess bandwidth available. Further, the frames are buffered in a packet delay buffer until an entire packet is received. Thus the frames have buffer that is more than the full packet level.

US20080056192 discloses a wireless broadband communications system and method that achieves reduced latency for high priority data when multiplexed with lower priority data for transmission over a TDD point-to-point radio link. The system prepares multiple data streams for transmission over a TDD radio link by buffering multiple data streams containing high and low priority packets in separate queues based upon their corresponding priority level. Each packet in the higher priority queues has a specified size, and a header defining the type of service provided and the packet destination. Next, the packets in the lower priority queues are fragmented to a reduced size based upon the data capacity of the link. However, the drawback is size of the fragmented frames cannot be chosen and although the wireless broadband communications system has reduced latency but consumes memory usage as it does not disclose about controlling errors.

Accordingly, there exists a need for a system and method for fragmenting jumbo Ethernet packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links including DWDM transmission system with reduced local memory usage and reduced processing transmission latency.

OBJECTS OF INVENTION
One or more problems present of the prior art may be overcome by various embodiment of the present invention.

Accordingly it is the primary object of the present invention to provide a system for fragmenting jumbo Ethernet frames or packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links including Dense Wavelength Division Multiplexing (DWDM) transmission system with reduced local memory usage and reduced processing transmission latency.

Another object of the present invention is to provide a method for fragmenting jumbo Ethernet frames or packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links.

SUMMARY OF INVENTION
Thus according to the basic aspect of the present invention there is provided a system for fragmenting jumbo Ethernet frames or packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links comprising of:
One or more Small Form factor Pluggable (SFP) optical transceivers;
Field Programmable Ball Grid Array (FPGA) memory module;
Optical Transport Network (OTN) mapper with Forward Error Correction (FEC) module;
Serializer/De-Serializer (SERDES) module; and
Dense Wavelength Division Multiplexing (DWDM) Transmitter/Receiver module,
wherein the FPGA memory module further comprising of:
One or more gigabit Ethernet physical (GigE PHY) devices and gigabit Ethernet Media Access Control (GigE MAC) devices;
Ethernet Receiver data (EthRxE) and Transmitter data (EthTxE) First in First out (FIFO) module;
Generic Framing Procedure Transmitter (GFPTxE) and Receiver (GFPRxE) module; and
SERDES Framer Interface Level 4 (SFI-4.1),
wherein the SFP optical transceiver connected to the FPGA memory module converts high speed serial electrical stream to a client optical wavelength and vice versa,
wherein the GigE physical device encodes /decodes GigE signal and converts the high speed serial stream to a low speed parallel stream,
wherein the low speed parallel stream is fed in to the GigE MAC block, which extracts the Ethernet frames from the parallel stream and performs non-intrusive monitoring of the Ethernet frames to check errors,
wherein the Ethernet frames received from one or more channels by the EthRxE are read from the memory and framed by the GFPTxE module and sent over the OTN mapper with FEC module,
wherein the OTN mapper with FEC module connected between the FPGA memory module and SERDES module control errors on data links where the EthTxE adds systematically generated error-correcting codes (ECC) to each Ethernet frame,
wherein the ECC allows the EthRxE to detect/correct a number of errors and avoid forcing the EthTxE to resend the Ethernet frame,
wherein the SFI-4.1 connects OTN framers and FEC processors to optical modules, and aggregates data bandwidths of GFP/Framing and net OTN bandwidth, and
wherein when the aggregate bandwidth after GFP/framing overhead is less than the net OTN bandwidth available, the EthTxE frame split into fragments during reception itself and sends over the GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module.

It is another aspect of the present invention, wherein the SERDES module converts the parallel stream coming from the OTN mapper with FEC module into a high speed serial stream depending on the client signal mode at which the OTN mapper with FEC module is configured.

It is another aspect of the present invention, wherein the high speed serial stream from the SERDES module is converted into DWDM grated optical signal by a DWDM OTN Framer and transponder in the DWDM transceiver module for long distance transmission.

It is another aspect of the present invention, wherein input/output of the OTN mapper with FEC module transmits through the SERDES module to the DWDM transceiver module.

It is another aspect of the present invention, wherein the EthTxE informs the GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module when fragment size of packets of data available or a packet end is received and the GFPTxE reads as soon as the fragmented packet arrived and sends over the OTN mapper with FEC module.

It is another aspect of the present invention, wherein the GFPRxE receives the framed packets and writes into specific channel EthTxE memory and informs on every full packet ending.

It is another aspect of the present invention, wherein the EthTxE transmits the Ethernet frames over Ethernet links and identifies the maximum packet and it’s transmit rate, to compute when to trigger its transmission before a full packet arrival thereby the EthTxE can have reduced buffer that is less than the full packet level and furthermore reduce the EthTxE latency to be less than a frame time.

Another aspect of the present invention is directed to provide a method for fragmenting jumbo Ethernet frames or packets to send over TDM/OTN links comprising steps of:
converting high speed serial electrical stream to a client optical wavelength and vice versa by SFP optical transceiver;
encoding / decoding of GigE signal and converting the high speed serial stream to a low speed parallel stream by GigE physical device;
extracting the Ethernet frames from the parallel stream and performing non-intrusive monitoring of the Ethernet frames to check errors;
receiving the Ethernet frames from one or more channels and copying into memory module;
framing the Ethernet frames from the memory module by a Generic Framing Procedure Transmitter (GFPTxE) module and sent over a Optical Transport Network (OTN) mapper with Forward Error Correction (FEC) module;
controlling errors on data links where the transmitter adds systematically generated error- correcting codes (ECC) to each Ethernet frame;
converting the parallel stream coming from the OTN mapper into a high speed serial stream depending on the client signal mode using a SERDES module; and
aggregating data bandwidths of GFP/Framing and net OTN bandwidth by SERDES Framer Interface Level 4 (SFI-4.1).

It is another aspect of the present invention, wherein the method further comprising steps of:
splitting EthTxE frames into fragments during reception itself and sending over the GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module; and
transmitting the Ethernet frames over Ethernet links and identifying maximum packet and transmit rate of the EthTxE when to trigger its transmission before a full packet arrival,
wherein the EthTxE frames are splitted into fragments when the aggregate bandwidth after GFP/framing overhead is less than the net OTN bandwidth available, and
wherein the EthTxE can have reduced buffer that is less than the full packet level and furthermore reduce the EthTxE latency to be less than a frame time.

It is another aspect of the present invention, wherein the fragmented size can be chosen depending on the excess bandwidth available over GFP/OTN channel.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWING:
Figure 1: illustrates the architecture of a system for fragmenting jumbo Ethernet packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links including Dense Wavelength Division Multiplexing (DWDM) transmission system.

DETAILED DESCRIPTION OF THE INVENTION WITH REFERENCE TO THE ACCOMPANYING FIGURE
The present invention is thus directed to a system for fragmenting jumbo Ethernet packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links including Dense Wavelength Division Multiplexing (DWDM) transmission system with reduced local memory usage and reduced processing transmission latency.

Referring to Figure 1, the system for fragmenting jumbo Ethernet frames or packets to send over TDM/OTN links comprises of one or more Small Form factor Pluggable (SFP) optical transceivers [1]; Field Programmable Ball Grid Array (FPGA) memory module [2]; Optical Transport Network (OTN) mapper with Forward Error Correction (FEC) module [3]; Serializer/De-Serializer (SERDES) module [4]; and Dense Wavelength Division Multiplexing (DWDM) Transmitter/Receiver module [5]. The SFP optical transceiver [1] is connected to the OTN mapper with FEC module [3] through the FPGA memory module [2]. The SFP optical transceiver [1] connected to the FPGA memory module [2] converts high speed serial electrical stream to a client optical wavelength and vice versa.

The OTN mapper with FEC module [3] connected in between the FPGA memory module [2] and SERDES module [4] is used to control errors on data links where the EthTxE [9] adds systematically generated error-correcting codes (ECC) to each frame, said ECC allows the EthRxE [8] to detect/correct a number of errors and avoid forcing the EthTxE [9] to resend the frame. The SERDES module [4] connected in between the OTN mapper with FEC module [3] and DWDM Transmitter/Receiver module [5] converts the parallel signal coming from the OTN mapper with FEC module [3] into a high speed serial stream depending on the client signal mode at which the OTN transponder is configured. The serial stream from the SERDES module [4] is fed into the DWDM OTN Framer and transponder in the DWDM Transmitter/Receiver module [5] which converts the high speed electrical stream into a DWDM grated optical signal for long distance transmission through the DWDM infrastructure and the input/output of the OTN mapper with FEC module [3] transmits through the SERDES [4] to the DWDM transceiver module [5].

The FPGA memory module [2] further comprises of one or more gigabit Ethernet physical (GigE PHY) devices [6] and gigabit Ethernet Media Access Control (GigE MAC) devices [7]; Ethernet Receiver data (EthRxE) [8] and Transmitter data (EthTxE) [9] First in First out (FIFO) module; Generic Framing Procedure (GFP) Transmitter (GFPTxE) [10] and Receiver (GFPRxE) module [11]; and SERDES Framer Interface Level 4 (SFI-4.1) [12]. The FPGA memory module [2] optionally includes Micro Processing Unit (MPU) [13].

The GigE physical devices [6] and [7] are connected to the Generic Framing Procedure (GFP) Transmitter (GFPTxE) [10] and Receiver (GFPRxE) module [11] through the Ethernet Receiver data (EthRxE) [8] and Transmitter data (EthTxE) [9] FIFO module. The GFP Transmitter (GFPTxE) [10] and Receiver (GFPRxE) module [11] is connected to the SFI-4.1 [12], said SFI-4.1 [12] is bi-directionally connected to the OTN mapper with FEC module [3]. The GigE physical device [6] performs physical layer encoding/decoding of the GigE signal and converts the high speed serial stream to a low speed parallel stream. The low speed parallel stream is fed into the GigE MAC block [7], which extracts the Ethernet frames from the parallel stream and performs non-intrusive monitoring of the Ethernet frames to check errors.

The Ethernet receiver engine (EthRxE) [8] receives Ethernet frames from one or more channels and are copied into the memory module. The Ethernet frames are read from the memory and are framed by the GFPTx Engine [10] and sent over the OTN mapper with FEC module [3]. The SFI-4.1 [12] is a SERDES based PHY specification used for connecting OTN framers and FEC processors to optical modules, and aggregates data bandwidths of GFP/Framing and net OTN bandwidth. When the aggregate bandwidth after GFP/framing overhead is less than the net OTN bandwidth available, the EthTxE [9] split into fragments during reception itself and sends over the GFP Transmitter (GFPTxE) [10] and Receiver (GFPRxE) [11] module (when GFP Transmitter (GFPTxE) [10] and Receiver (GFPRxE) [11] module is channelized and each GFP channel Bandwidth is more than Ethernet arrival rate). The EthTxE [9] informs the GFP Transmitter (GFPTxE) [10] and Receiver (GFPRxE) [11] module when fragment size of packets of data available or a packet end is received and the GFPTxE [10] reads as soon as the fragmented packet arrived and sends over the OTN mapper with FEC module [3].
The GFPRxE [11] receives the framed packets and writes into specific channel EthTxE [8] memory and informs on every full packet ending. The EthTxE [9] transmits the Ethernet frames over Ethernet links and identifies the maximum packet and it’s transmit rate, to compute when to trigger its transmission before a full packet arrival thereby the EthTxE [9] can have reduced buffer that is less than the full packet level and furthermore reduce the EthTxE [9] latency to be less than a frame time.

In another aspect, a method for fragmenting jumbo Ethernet frames or packets to send over TDM/OTN links comprising steps of converting high speed serial electrical stream to a client optical wavelength and vice versa by SFP optical transceiver; performing physical layer encoding / decoding of GigE signal and converting the high speed serial stream to a low speed parallel stream by GigE physical device; extracting the Ethernet frames from the parallel stream and performing non-intrusive monitoring of the Ethernet frames to check errors; receiving the Ethernet frames from one or more channels and copying into memory module; framing the Ethernet frames from the memory module by a Generic Framing Procedure Transmitter (GFPTxE) module and sent over a Optical Transport Network (OTN) mapper with Forward Error Correction (FEC) module; controlling errors on data links where the transmitter adds systematically generated error-correcting codes (ECC) to each Ethernet frame; converting the parallel stream coming from the OTN mapper into a high speed serial stream depending on the client signal mode using a SERDES module; and aggregating data bandwidths of GFP/Framing and net OTN bandwidth by SERDES Framer Interface Level 4 (SFI-4.1).

The method for fragmenting jumbo Ethernet frames or packets to send over TDM/OTN links further comprising steps of splitting EthTxE frames into fragments during reception itself and sending over the GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module; and transmitting the Ethernet frames over Ethernet links and identifying maximum packet and transmit rate of the EthTxE when to trigger its transmission before a full packet arrival. The EthTxE frames are splitted into fragments when the aggregate bandwidth after GFP/framing overhead is less than the net OTN bandwidth available. The fragmented size can be chosen depending on the excess bandwidth available over GFP/OTN channel. The EthTxE can have reduced buffer that is less than the full packet level and furthermore reduce the EthTxE latency to be less than a frame time.

For illustration:
When the aggregate bandwidth after GFP/Framing overhead is less than the net OTN bandwidth available, the EthTxE Frames split into fragments during reception itself and sent over GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module (assuming GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module is channelized and each GFP channel Bandwidth is more than Ethernet arrival rate). The fragmented size can be chosen to be multiple of 2^n for hardware implementation. For experimentation 128 bytes was chosen however 64, 256 or 512 or more can be chosen as required. The choice depends on the excess bandwidth available over GFP/OTN channel.
As the fragmented size reduces, the net overhead is much more. The fragment size can be chosen to be smallest Ethernet frame so that all Ethernet frames are split to the level smallest Ethernet frames and sent over. This size can be like 64 bytes. The EthTxE informs the GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module when the fragment size of packet size of data available or a packet end is received. The GFPTxE reads as soon as fragmented packet arrived and sends over the OTN mapper with FEC module. The EthTxE needs to have memory that is close to 128 bytes or close the fragment size chosen.

Also the latency on GFPTxE side reduced to only 128 bytes time rather than full frame time (which are significant in case of Jumbo frames). The GFPRxE takes the partial frames and writes into EthTxE buffer and informs on every full packet ending. The EthTxE identifies the maximum packet and its transmitting rate, to compute when to trigger its transmission before a full packet arrival. Thus the EthTxE can have reduced buffer that is less than the full packet level and also reduce the EthTxE latency to be less than a frame time according to the present invention.

The present invention may be embodied in other specific configuration without departing from its spirit or essential characteristics. The described embodiment is to be considered in all respects only as illustrative and not restrictive.

,CLAIMS:WE CLAIM:
1. A system for fragmenting jumbo Ethernet frames or packets to send over Time Division Multiplexing (TDM)/Optical Transport Network (OTN) links comprising of:
One or more Small Form factor Pluggable (SFP) optical transceivers [1];
Field Programmable Ball Grid Array (FPGA) memory module [2];
Optical Transport Network (OTN) mapper with Forward Error Correction (FEC) module [3];
Serializer/De-Serializer (SERDES) module [4]; and
Dense Wavelength Division Multiplexing (DWDM) Transmitter/Receiver module [5],
wherein the FPGA memory module [2] further comprising of:
One or more gigabit Ethernet physical (GigE PHY) devices [6] and gigabit Ethernet Media Access Control (GigE MAC) devices [7];
Ethernet Receiver data (EthRxE) [8] and Transmitter data (EthTxE) [9] First in First out (FIFO) module;
Generic Framing Procedure Transmitter (GFPTxE) [10] and Receiver (GFPRxE) [11] module; and
SERDES Framer Interface Level 4 (SFI-4.1) [12],
wherein the SFP optical transceiver [1] connected to the FPGA memory module [2] converts high speed serial electrical stream to a client optical wavelength and vice versa,
wherein the GigE physical device [6] encodes /decodes GigE signal and converts the high speed serial stream to a low speed parallel stream,
wherein the low speed parallel stream is fed in to the GigE MAC block [7], which extracts the Ethernet frames from the parallel stream and performs non-intrusive monitoring of the Ethernet frames to check errors,
wherein the Ethernet frames received from one or more channels by the EthRxE [8] are read from the memory and framed by the GFPTxE module [10] and sent over the OTN mapper with FEC module [3],
wherein the OTN mapper with FEC module [3] connected between the FPGA memory module [2] and SERDES module [4] control errors on data links where the EthTxE [9] adds systematically generated error-correcting codes (ECC) to each Ethernet frame,
wherein the ECC allows the EthRxE [8] to detect/correct a number of errors and avoid forcing the EthTxE [9] to resend the Ethernet frame,
wherein the SFI-4.1 [12] connects OTN framers and FEC processors to optical modules, and aggregates data bandwidths of GFP/Framing and net OTN bandwidth, and
wherein when the aggregate bandwidth after GFP/framing overhead is less than the net OTN bandwidth available, the EthTxE [9] frame split into fragments during reception itself and sends over the GFP Transmitter (GFPTxE) [10] and Receiver (GFPRxE) [11] module.

2. The system as claimed in claim1, wherein the SERDES module [4] converts the parallel stream coming from the OTN mapper with FEC module [3] into a high speed serial stream depending on the client signal mode at which the OTN mapper with FEC module [3] is configured.

3. The system as claimed in claim 2, wherein the high speed serial stream from the SERDES module [4] is converted into DWDM grated optical signal by a DWDM OTN Framer and transponder in the DWDM transceiver module [5] for long distance transmission.

4. The system as claimed in claim 3, wherein input/output of the OTN mapper with FEC module [3] transmits through the SERDES module [4] to the DWDM transceiver module [5].

5. The system as claimed in claim 1, wherein the EthTxE [9] informs the GFP Transmitter (GFPTxE) [10] and Receiver (GFPRxE) [11] module when fragment size of packets of data available or a packet end is received and the GFPTxE [10] reads as soon as the fragmented packet arrived and sends over the OTN mapper with FEC module [3].

6. The system as claimed in claim 5, wherein the GFPRxE [11] receives the framed packets and writes into specific channel EthTxE [9] memory and informs on every full packet ending.

7. The system as claimed in claim 6, wherein the EthTxE [9] transmits the Ethernet frames over Ethernet links and identifies the maximum packet and it’s transmit rate, to compute when to trigger its transmission before a full packet arrival thereby the EthTxE [9] can have reduced buffer that is less than the full packet level and furthermore reduce the EthTxE [9] latency to be less than a frame time.

8. A method for fragmenting jumbo Ethernet frames or packets to send over TDM/OTN links comprising steps of:
converting high speed serial electrical stream to a client optical wavelength and vice versa by SFP optical transceiver;
encoding / decoding of GigE signal and converting the high speed serial stream to a low speed parallel stream by GigE physical device;
extracting the Ethernet frames from the parallel stream and performing non-intrusive monitoring of the Ethernet frames to check errors;
receiving the Ethernet frames from one or more channels and copying into memory module;
framing the Ethernet frames from the memory module by a Generic Framing Procedure Transmitter (GFPTxE) module and sent over a Optical Transport Network (OTN) mapper with Forward Error Correction (FEC) module;
controlling errors on data links where the transmitter adds systematically generated error- correcting codes (ECC) to each Ethernet frame;
converting the parallel stream coming from the OTN mapper into a high speed serial stream depending on the client signal mode using a SERDES module; and
aggregating data bandwidths of GFP/Framing and net OTN bandwidth by SERDES Framer Interface Level 4 (SFI-4.1).

9. The method as claimed in claim 8 further comprising steps of:
splitting EthTxE frames into fragments during reception itself and sending over the GFP Transmitter (GFPTxE) and Receiver (GFPRxE) module; and
transmitting the Ethernet frames over Ethernet links and identifying maximum packet and transmit rate of the EthTxE when to trigger its transmission before a full packet arrival,
wherein the EthTxE frames are splitted into fragments when the aggregate bandwidth after GFP/framing overhead is less than the net OTN bandwidth available, and
wherein the EthTxE can have reduced buffer that is less than the full packet level and furthermore reduce the EthTxE latency to be less than a frame time.

10. The method as claimed in claim 9, wherein the fragmented size can be chosen depending on the excess bandwidth available over GFP/OTN channel.