1. Field of the Invention
This invention relates to the field of data communications, and more particularly to increasing data transmission reliability for Fixed Wireless Applications.
2. Description of the Related Art
Data comniunications network and device designers are continually striving to increase capacity, increase throughput, improve efficiency and improve overall operations according to need. As such, many use specific designs are being pursued that are appropriate for particular operations. For example, an entire fixed wireless industry has developed for providing wireless service to relative non-mobile users. Heretofore, the fixed wireless industry has utilized known cellular standards such is IS-95 code division multiple access protocols and technology for delivering fixed wireless cellular service to the subscriber. The fixed wireless service was originally implemented as a carrier service, initially intended for broadcast of television, and has commonly been referred to as wireless cable. In the United States, this service is available at 2.5 GHz. The service is usually analog and one way (transmit only), with a range of about 30 miles (50 km), and has been deployed in the United States, the Middle East, Latin America, Eastern Europe and Asia Pacific. Carriers have not been successful with television programming and have instead been able to offer two-way service for Internet. Others offer Internet service by using a hybrid approach with a PSTN (public switched telephone network) connection for the return path. Two-way service reduces the effective range of MMDS to about 6 miles (10 km).
Since the inception of fixed wireless networks, however, digital subscriber lines, cable modems and other broadband access technologies have emerged enabling their users to achieve fast internet service in relation to dial up subscriber speeds. The Data-Over-Cable Service Interface Specification (DOCSIS) was defined to promote interoperability among cable modems and other related products. It has been modified by the Broadband Wireless Internet Forum (BWIF), however, to be used in fixed wireless point to multipoint distribution services using physical layers such as orthogonal frequency division multiplexing (OFDM). Because wireless signals are more susceptible to channel problems such as fading, it is desirable to make the wireless data flow more robust to account for transmission errors and interference. Thus, there exists a need for a DOCSIS based fixed wireless system that provides protection for the transmission errors and interference.
An apparatus and method of the present invention includes implementing an automatic repeat request (ARQ) protocol to improve the robustness of fixed wireless and other wireless communication systems. Because some systems have adopted an approach of utilizing an adapted DOCSIS protocol for wireless applications, the present invention contemplates modifying the adapted DOCSIS protocol by introducing an ARQ protocol that functions in a DOCSIS wireless environment. More specifically, an ARQ protocol is implemented at the medium access control (MAC) layer to provide fast response relative to conventional ARQ protocols and logic that are implemented at higher layers (e.g., TCP/IP) in the OSI layered approach to system architecture and design. To achieve this result, a DOCSIS defined signal and header is modified to include an ARQ header containing ARQ parameters. While some known approaches to ARQ involve retransmission of entire frames or blocks, the present invention contemplates provide control down to the packet data unit (PDU) level in some embodiments. To achieve this level of granularity, the ARQ header includes sequence numbers within each ARQ header on a packet data unit (PDU) basis. Accordingly, when a receiver, by evaluating sequence numbers for received PDUs determines that it did not received a particular PDU, the receiver generates a “non-acknowledge” signal to the transmitter to advise it of the failure and to identify the missing PDU. As the transmitter includes a buffer that stores all transmitted PDUs for a specified period, the transmitter is able to find the missing or lost PDU and retransmit the missing PDU to the receiver.
The inventive method and apparatus disclosed herein are particularly advantageous in that they provide a capability for transmitting wireless communication signals in a more robust manner than is provided for, especially for particular networks utilizing protocols such as cable modem protocols that have been adopted for wireless communications. More specifically, the invention makes use of the features of a modified DOCSIS MAC layer to provide the ARQ protocol for fixed wireless applications employing the DOCSIS standard as originated for cable modems but modified for fixed wireless transmissions. In one embodiment of the invention, a selective repeat ARQ is utilized, with negative acknowledge (N-ACK) on all packages, and, more particularly, as a MAC based ARQ for the packages. Here, the ARQ operates on a per PDU basis at the MAC layer. Control is accomplished through messaging. Preferably, the ARQ protocol can be enabled either on a connection or on a per flow basis. Negative acknowledge includes the receiver advising the transmitter when a PDU was not received. When the transmitter has received a negative acknowledge signal, it deletes or flushes all PDUs transmitted prior to the PDU identified in the negative acknowledge signal. Additionally, a transmitter flushes its buffers if, after a specified period, it has not received a N-ACK messag.
One embodiment of the invention is implemented at the MAC layer because it is superior in performance and efficiency over implementing the ARQ protocol on the TC/IP layer or the physical layer. Implementing the ARQ protocol on the TC/IP layer would be very slow and inefficient. In a down stream broadcast environment, a physical layer ARQ would swamp the upstream channel because the N-ACK would have to be issued on a per OFDM burst basis, which is typically only about 64 bytes. Moreover, each of the receiving physical layers have no information regarding whether a missed burst was intended for them, and thus will always notify the head end that the burst was missed.
If one assumes a reasonable channel, a selective repeat is used to minimize data amounts and latency in retransmission. The negative or non-acknowledge signal (N-ACK) minimizes the signaling overhead for each packet. An upstream ARQ improves packet delivery from signal unit (SU) 10 to head end (HE) 20 (
Upstream ARQ improves packet delivery from SU to HE, while downstream ARQ improves packet delivery from HE to SU. The logic is the same in both the upstream and downstream paths.
General Protocol Features are as Follows:
The transmitter adds an ARQ header to the frames and keeps a copy of the transmitted frames in a transmitter window until it either receives a N-ACK or it times out and assumes the packet was received correctly. There are at least three timers utilized in the described embodiment of the invention. They are:
Additionally, the invention includes at least two counters. They are:
Generally, the receiver confirms received frames and notifies missing sequence numbers via the negative acknowledgement messages that are sent based on a periodic receiver time or on an explicit acknowledgement request from the transmitter.
The receiver buffers out-of-sequence frames until in-sequence delivery is provided. The transmitter retransmits only missing frames requested by the receiver. The transmitter can send an explicit acknowledgement request based on a transmitter timeout.
If the sequence number is not valid, all the packets in the current window will be flushed and will be set to the sequence number of the packet received. When a N-ACK message is received all packets up to the sequence number are released, because this means that packets up to that sequence number are no longer live at the transmitter side.
When a Request_NACK message is received, the sequence number field is read from the Request_NACK message. All packets from up to that sequence number are released. The packets in the new receive window (after releases) are examined and the sequence number of all missing packets is constructed. An N-ACK message with the latest missing sequence numbers is then immediately sent, and all packets from Fr up to and including Lr,init are flushed.
Three ARQ control frame formats are proposed as described below. The ARQ can be implemented using either MAC Management Messages or EHDR. The wireless access termination system (WATS) and wireless modem (WM) capabilities are negotiated during the setup of an ARQ session. An ARQ header for control frame is inserted as EHDR header as in the ARQ data frame format. A reserved EH type=8 is allocated for ARQ. For the ARQ header, an operation bit is used to indicate ARQ. The control bits indicate this is an ARQ control frame for Request_ACK message or ACK message. Since EH_LEN field is 4 bits, it limits the size of the bitmap sent to be 104 for ARQ per SU, and 72 for ARQ per flow.
While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof have been shown by way of example in the drawings and detailed description. It should be understood, however, that the drawings and detailed description thereto are not intended to limit the invention to the particular form disclosed, but on the contrary, the invention is to cover all modifications, equivalents and alternatives falling within the spirit and scope of the present invention as defined by the claims. As may be seen, the described embodiments may be modified in many different ways without departing from the scope or teachings of the invention.
This application incorporates by reference the following U.S. Non-Provisional Patent Applications each of which includes at least one of Anders Hebsgaard or Mark Dale as inventor: (1) U.S. patent application Ser. No. 09/430,821, filed Oct. 29, 1999, by Hebsgaard, et al., and incorporated herein by reference in its entirety; (2) U.S. patent application Ser. No. 09/574,558, filed May 19, 2000, by Hebsgaard, et al., and incorporated herein by reference in its entirety; (3) U.S. patent application Ser. No. 09/653,155, filed Aug. 31, 2000, by Hebsgaard, et al., and incorporated herein by reference in its entirety; and (4) U.S. patent application Ser. No. 10/097,942, filed Mar. 15, 2002, by Dale, et al., and incorporated herein by reference in its entirety;