EP1364496A4 - Dynamic bandwidth allocation - Google Patents

Abstract

A method and apparatus for managing measured burst profile characteristics (630) for each of a plurality of remote sites which are dynamically allocated to one of a plurality of time slots (620) within a frame of a TDM wireless communication system containing at least one identified antenna (610) and at least one identified frequency (615) is disclosed. The method comprises the steps of determining a burst profile table (630) to store burst characteristic data (635), determining within the burst profile table (630), a location associated with a corresponding one of the remote sites, where the location is related to an identifier unique to the remote site contained within a corresponding time slot (625), and storing a corresponding measured burst profile (635) at the determinated location. In the aspect of the invention, management of a multiple-antenna (610), multiple-frequency (615) configuration includes providing known relations between antenna identifiers (610) to determine locations for storing burst profile characteristic data (635) for corresponding remote sites.

Description

DYNAMIC BANDWIDTH ALLOCATION

Field of the Invention

Cross Reference to Related Applications

[0001] This application claims priority of provisional application number 60/266,475, filed on February 6, 2001, entitled "Software Provisional Application", which hereby incorporated by reference in its entirety.

[0002] The present invention relates generally to wireless broadband technology and more particularly to dynamic allocation of bandwidth for point-to-multipoint systems.

Background

[0003] Point to multipoint (PTMP) systems, such as wireless communication systems are constantly being demanded to accommodate more channels. This requires more efficient allocation of available limited bandwidth and scheduling of system resources and capacities. Wireless communication systems facilitate two-way communication between a plurality of subscriber radio stations or subscriber units (referred to as hub stations and remote stations). Example systems include mobile cellular telephone systems and personal communication systems (PCS). The objective of many wireless communication systems is to provide communication channels on demand between the hub stations and the remote stations. In the wireless systems using multiple access schemes, frames of time are the basic transmission unit. Each frame is divided into a plurality of slots of time. Some time slots are used for control purposes and some time slots are used for information transfer. Information is typically transmitted during time slots in the frame, in which the time slots are specifically assigned. Transmissions from the hub station to the remote stations are often referred to as forward link transmissions. Transmissions from the remote stations to the hub station are commonly referred to as reverse link transmissions.

[0004] Wireless communication systems typically use either time division duplexing

(TDD) or frequency division duplexing (FDD) methods to facilitate the exchange of information between the hub station and the remote stations. Both the TDD and FDD duplexing schemes are well known in the art. In FDD systems, duplexing of transmissions between a hub station and its remote stations is performed in the frequency domain. Different sets of frequencies are allocated for forward and reverse transmissions. In TDD systems, duplexing of transmissions between a hub station and its remote stations is performed in the time domain. The channel is time-divided into repetitive time periods or time "slots" which are employed for forward and reverse transmissions.

[0005] Bandwidth requirements for PTMP systems vary as a function of time. For example, in PTMP systems offering broadband services the forward link and reverse link transmissions may have unequal or asymmetrical bandwidth requirements. Also, the ratio of forward link/reverse link bandwidth desired may vary between stations and/or between channels. Further, the need for symmetrical or asymmetrical communication in a channel varies depending upon the type of user/remote station. Thus, there is a need for a system and method that can dynamically and adaptively allocate available bandwidth.

[0006] Furthermore, system resources/capacities should be efficiently scheduled to accommodate the increased demand for channels. For example, many systems transmit information in accordance with priorities assigned to the information/data. That is, the highest priority data will be scheduled for transmission first before lower priority data is scheduled. A disadvantage of such a scheduling scheme is during times of heavy traffic congestion, lower priority lower priority data may not be scheduled within an allow amount of time if an abundance of highest priority data is to be scheduled. Another scheme is to schedule on a first in first out (FIFO) basis. However, this may lead to highest priority data not being scheduled/transmitted with an allowable amount time. Another disadvantage of current systems is they tend to have an inherent predetermined slant toward favoring the forward link transmissions over reverse link transmission. Thus, there is a need for a system and method for dynamically and adaptively scheduling resources/capacities and allocating bandwidth of systems such as PTMP systems.

Summary of the Invention

[0007] A technique for allocating bandwidth for a communications path of a point to multipoint (PTMP) system having a plurality of communications paths allocates available bandwidth within each communications path in accordance with a predetermined priority and an arrival time of information to be transmitted within the available bandwidth.

Brief Description of the Drawings

[0008] The above and other advantages and features of the present invention will be better understood from the following detailed description of the preferred embodiments of the invention, which is provided in connection with the accompanying drawings. The various features of the drawings may not be to scale. Included in the drawing are the following figures:

[0009] Figure 1 is a stylized illustration of a PTMP system;

[0010] Figure 2 is an illustration of time division formatting in accordance with an embodiment of the present invention;

[0011] Figure 3 is a diagram illustrating the relationship between hub stations, remote stations, and communication paths in accordance with an embodiment of the present invention;

[0012] Figure 4 is functional block diagram of a real time adaptive scheduling and bandwidth allocation system, in accordance with an embodiment of the present invention;

[0013] Figure 5 is an illustration depicting centralized scheduling priority inversion;

[0014] Figure 6 is a functional block diagram of an exemplary inter-path scheduler in accordance with the present invention; and

[0015] Figure 7 is a flow diagram of an exemplary process for scheduling capacity and allocating bandwidth in accordance with an embodiment of the present invention.

Detailed Description

[0016] A capacity scheduling and bandwidth allocation system and method as described herein utilize centralized and distributed capacity scheduling to accomplish real time adaptive capacity scheduling and bandwidth allocation within a wireless point to multipoint (PTMP) environment. Figure 1 is a stylized illustration of a PTMP system 100. System 100 comprises at least one hub station 24 and at least one remote station 22. A remote station 22 may comprise a local area network (LAN), such as depicted by LANs 26 and 28, individual processors, such as depicted by processor 30, a wireless communication means, such as facilitated by and depicted by antenna 32, an electromagnetic and/or optic communication means, such as facilitated by and depicted by conductor 34, or any combination thereof. A hub station 24 may comprise a processor, such as depicted by processor 36, a wireless transmission means, such as facilitated by and depicted by antenna 38, an electromagnetic and/or optic communication means, such as facilitated by and depicted by conductor 34, or any combination thereof. In an exemplary embodiment, antenna 38 comprises an omnidirectional antenna for achieving concurrent communication with all other stations. It is understood however, that antenna 38 may comprise a plurality of directional antennas. As shown by system 100, communication may be achieved between a plurality of hub stations 24, between hub stations 24 and remote stations 22, between a plurality of remote stations 22, or any combination thereof.

[0017] Of particular interest are time-division multiplexing (TDM) and time-division multiple access (TDMA) systems, wherein communication between stations is accomplished by formatting information into unique time segments. A description of TDM and TDMA systems may be found in any of several textbooks pertaining to wireless communications, such as "Wireless Communications: Tdma versus Cdma", by Savo G. Glisic and Pentti A. Leppanen, June 1997; and "Wireless Communications & Networks", by William Stallings, August 23, 2001; for example. Figure 2 is an illustration of time division formatting in accordance with the present invention. As shown in Figure 2, a communication channel comprises a plurality of frames 42. Each frame 42 is partitioned into a forward link portion 44 and a reverse link portion 46. During the forward link 44, information is transmitted from a hub station (e.g., hub station 24) to at least one remote station (e.g., remote station 22). During a reverse link 46, information is transmitted from the remote station(s) (e.g., remote station 22) to the corresponding hub station (e.g., hub station 24). Each of the forward link portions 44 and the reverse link portions 46 is divided into a plurality of time slots 48. In accordance with the scheduling and bandwidth allocation system and method described herein, the length of each time slot 48 is separately and dynamically reconfigurable in accordance with the demand for resources, including bandwidth, placed on the PTMP system (e.g., system 100). [0018] Figure 3 is a diagram illustrating the relationship between hub stations 24, remote stations 22, and communication paths (CPs). Communication between each remote station 22 and a hub station 24 is accomplished via a communication path (CP). As shown in Figure 3, and described with reference to Figure 1, communication between stations may be accomplished via several types of media, such as an airlink, a vacuum, an electrical conductor, an optical conductor, or any combination thereof. In accordance with the system and method described herein, multiple communication paths (CPs) share and utilize the limited number of physical time slots typical in wireless TDM or TDMA architectures. These architectures may comprise various time division schemes, such as frequency division duplexing (FDD), time division duplexing (TDD), or any combination thereof, for example. A single CP is associated with communication between a hub station 24 and each remote station 22 of a wireless PTMP environment (e.g., system 100). A single CP is capable of supporting multiple application sessions concurrently, such as multiple Internet Protocol (IP) packet flows, and asynchronous transfer mode (ATM) virtual connections, for example. The exemplary configuration shown in Figure 3 comprises an ATM application utilizing 128 CPs. Each CP comprises a plurality of forward and reverse time divisions 48. Each ATM access point comprises N virtual connection (VC) terminations. It is understood that this configuration depicted in Figure 3 is exemplary. For example, the number of CPs may be more or less than 128, the number of VCs associated with each multiplexer (MUX) may differ (i.e., not all equal to N), and the application may be other than asynchronous transfer mode (ATM), such as Internet Protocol (IP), precedence service and recent IP quality of service (QOS) enhancements, open systems interconnection (OSI), and multi protocol label switching (MPLS), for example.

[0019] Information to be communicated is segmented into fixed sized data blocks called

Protocol Data Units (PDUs). Each PDU comprises a payload portion, and a prepended portion. The payload portion contains data indicative of the information to be transmitted and/or received. The prepended portion contains data indicative of the priority of the information and the amount of time a PDU has been waiting to be scheduled. As explained in more detail herein, the data in the prepended portion are updated dynamically and adaptively to provide scheduling and bandwidth allocation. To accommodate all the information to be communicated in a PTMP system, such as system 100, the PDUs are scheduled in accordance with priorities assigned to the information, the amount of time a PDU has been waiting to be transmitted, and parameters associated with each CP. Information is assigned a priority, or multiple priorities, in accordance with a specific class-of-service. A class-of-service (COS) is a designation used to describe the service treatment and privileges given to a particular entity, such as a terminal or CP. For example, the highest priority may be assigned to information having a "time sensitive" COS, and lessor priorities may be assigned to information having either "not time sensitive" or "best effort" classes-of-service.

[0020] Each CP supports a multiple, but limited, number of priority based classes-of- service (COS). Each CP is associated with parameters contained in a respective traffic management contract (TMC). Thus, in accordance with the real time dynamic scheduler describe herein, each CP may have its own unique TMC, all CPs may use the same TMC, or any combination thereof. Each TMC comprises parameters, which are utilized by the real time dynamic scheduling system to allocate capacity resources and bandwidth, and to select PDUs for transmission. A TMC is capable of being updated in real time. Each TMC contains parameters such as expected minimum and peak asymmetric transmission rates, maximum allowable delay increment for each class-of-service (COS), and discard delay threshold per COS, for example. In one embodiment of the invention, each TMC comprises the attributes as described in Table 1.

Table 1 Communication Path Traffic-Management Contract (CP TMC)

[0021] The above-described CP COS and TMC system and method are readily applicable to a plurality of wireless communications systems. For example, the various ATM based quality of service (QOS) offerings may be grouped and/or mapped onto the CP priority based COS, and each CP's TMC may be updated in real time to reflect the sum of the minimum and peak ATM virtual connection (VC) rates that it supports. Furthermore, because a TMC can be updated in real time, the transmission of both ATM permanent virtual connections (PVC) and switched virtual connections (SVC) may be supported.

[0022] Referring to Figure 4, there is shown a functional block diagram of a real time adaptive scheduling and bandwidth allocation system. Capacity scheduling and bandwidth allocation utilizes inter-path scheduling and intra-path scheduling. Inter-path scheduling aids in determining the allocation of available capacity between multiple competing communications paths (CPs). Intra-path scheduling aids in determining PDU transmission within the context of a single CP direction.

[0023] The exemplary block diagram shown in Figure 4 illustrates scheduling for two

CPs. The inter-path scheduler 52 distributes bulk capacity grants to the forward path intra-path scheduler 54(a) and the reverse path intra-path scheduler 54(b). A bulk capacity grant comprises information indicative of the explicit number of PDUs, which may be transmitted in each direction within a specified transmission frame (also referred to as an airlink frame herein). This information is distributed to the remote stations 22 via airlink control channel signaling. In an exemplary embodiment, this information is distributed to remote stations 22 one airlink frame time in advance of actual implementation.

[0024] During the forward portion of the transmission frame, each active instance of forward intra-path scheduling utilizes the bulk capacity grant provided by the inter-path scheduler 52 to make explicit PDU selections among the PDUs within its COS queuing mechanism. Intra-path scheduling utilizes an instantaneous priority calculation to make explicit PDU selections and then in turn generates a capacity report detailing the quantities of PDUs remaining queued within the CP for the given direction. This capacity report is provided as input into the path's bid processor 58.

[0025] During the reverse portion of the airlink frame, each active instance of reverse intra-path scheduling utilizes the bulk capacity grant provided by the inter-path scheduler 52 to make explicit PDU selections among the PDUs within its COS queuing mechanism. The intra- path scheduler 54 utilizes an instantaneous priority calculation to make explicit PDU selections and then in turn generates a capacity report detailing the quantities of PDUs remaining queued within the CP for the given direction. This capacity report is provided in an airlink reverse control-channel and is input into the path's bid processor 58. [0026] The path's bid processor 58 utilizes the capacity reports provided by the intra-path scheduler 54 to determine newly arrived PDUs that require consideration by the inter-path scheduler 52. These newly arriving PDUs are indicated to the inter-path scheduler 52. The inter- path scheduler 52 then utilizes the data associated with the newly arrived PDUs to redistribute the next bulk capacity grant, thus ensuring a system wide closed loop scheduling mechanism.

[0027] Two instances or intra-path scheduling are implemented for each wireless CP. A first instance for the forward direction (hub station 24 to remote station 22) and the other instance for the reverse direction (remote station 22 to hub station 24). The distributed intra-path scheduler 54 prevents premature priority inversion and incremental delays of protocol data units (PDUs) due to centralized scheduling latency. For example, consider the exemplary situation illustrated in Figure 5. As shown in Figure 5, a priority 1 PDU has arrived at a remote location 22 after a priority 2 PDU has been previously bidded and granted capacity. When the remote station 22 receives its grant of one (1) PDU, it is faced with a scheduling dilemma. It now has two PDUs to transmit and the one it has been granted capacity for is a lower priority and less delay sensitive than the later arriving, higher priority 1 PDU.

[0028] An adaptive real time scheduler as described herein tends to prevent premature priority inversion and mitigate its adverse effects resulting from incremental delays of PDU transmission due to scheduling latency. The remote station 22 has the capability to ultimately decide which PDU to transmit in accordance with the bulk capacity grant. Thus, the priority 1 PDU of Figure 5 may be transmitted before the priority 2 PDU. Accordingly, the hub station (24) based inter-path scheduler 52 adapts its real time scheduling behavior to accommodate the distributed decisions being made by the intra-path schedulers 54.

[0029] The intra-path scheduler 54 is a distributed dynamic priority based algorithm.

The intra-path scheduler 54 schedules PDUs utilizing priority, while allowing controlled priority inversion when dictated by the parameters of the path's TMC. This attribute of the intra-path scheduler is referred to as dynamic delay-weighted prioritization (DDWP). The inter-path scheduler 54 performs prioritization based scheduling in accordance with delay sensitivities of the PDUs involved. The DDWP (intra-path) scheduler 54 comprises the capability to determine which PDU(s) will be transmitted in accordance with a bulk capacity grant from the associated inter-path scheduler 52. [0030] The intra-path scheduler 54, in accordance with DDWP, determines and assigns an instantaneous priority for each PDU at the front of each class-of-service transmission queue. This mechanism utilizes a time dependent instantaneous priority index that is then used in explicit PDU selections. The instantaneous priority for each PDU is determined in accordance with the following equation:

P_j(t) = MDI_j - ω,(t), (1)

where, t represents time; P_j(t) represents an instantaneous priority for a j_th class-of-service as a function of time; MDI_j represents a communication path's maximum allowable delay increment for a j_th class-of-service; and co,(t) represents a value indicative of the longest amount of time a PDU of a j_tll class-of-service has been waiting for allocation. This instantaneous priority index, P,(t), is utilized to determine explicit inter-class PDU selection. Thus, if delay has increased within a lower priority PDU stream of class 'j' to the point that it exceeds the path's maximum allowable delay, MDI,, for that class-of-service, and a degree of freedom exists within the higher priority PDU stream 'j-1' (i.e., class 'j-F will not exceed its MDI^._!) given a priority inversion), DDWP dictates that the lower priority PDU is selected over the higher priority PDU. It is to be understood, a real time adaptive scheduler and bandwidth allocation system and method as described herein is applicable to a system comprising any number of classes of service are applicable.

[0031] Once the intra-path scheduler 52 has determined which explicit PDUs it will utilize in fulfilling its bulk capacity grant, the CP is responsible for updating the inter-path scheduling function regarding any newly arrived PDU capacities that require scheduled capacity. To accomplish this, the hub resident CP capacity bid processor 58 keeps a history of the path's PDU capacities per class-of-service. This history comprises counts of PDUs remaining queued within the CP the previous time it reported its queue capacities. These counts also include the counts of PDUs that have been granted by the inter-path scheduler 52 since the most recent capacity update. Thus, utilizing the actual capacities reported by the intra-path scheduling instances, the CP's capacity bid function calculates any newly arriving capacities and reports these to the inter-path scheduler 52. Newly arriving PDUs for each class-of-service are calculated as follows: PINEW = PI ACTUA + PIGRANT - PIPREV, P2NEW = P2_ACTUAL + P2GRANT - P2_PREV. etc. (PI represents priority 1, P2 represents priority 2, etc.) By assuming the number of PDUs granted as transmitted by the intra-path scheduling instances, the hub based capacity bid function continually adjusts to the actual reported queue quantities of the intra-path scheduling function. Thus by doing so, it continually synchronizes the centralized inter-path scheduling function to the distributed intra-path decision making processes. After providing the newly arriving PDU counts to the inter-path scheduling function PΠPREV becomes PΠ_ACTUAL and PΠ_GRANT becomes PΠ_NEW. thus the path's bid function is rearmed for the next capacity report.

[0032] A real time adaptive scheduling and bandwidth allocation system and method as described herein comprises one or more independent instances of inter-path scheduling. In one embodiment, these instances are located at a hub station 24. These inter-path scheduling points may be one per carrier, one per carrier group, one per hub, or one instance may exist for a group of hubs. The inter-path scheduler 52 determines the asymmetric bulk capacity grants per CP under its oversight and is responsible for maintaining fairness among competing CPs. In any of the above mentioned configurations, there is a one to many relationship (e.g., a 1:N relationship) between inter-path scheduling and intra-path scheduling as shown in Figure 6.

[0033] Figure 6 is a functional block diagram of an exemplary inter-path scheduler.

Intra-path capacity requests are submitted to the inter-path scheduler 52 in the form of PDU capacities. Each PDU submitted for scheduling is then assigned a priority indicator and queued in priority order within the appropriate class-of-service (COS) schedule. A priority indicator sort key is determined by the inter-path's Windowed Fair Priority Queuing (WFPQ) processor 62, which is used in determining CP and asymmetric fairness within each COS. WFPQ is described in more detail below. The inter-path scheduler 52 utilizes a simple, fixed-priority processing scheme and converts the queued PDU schedules into CP bulk capacity grants each scheduling interval. As depicted in Figure 6, during each scheduling interval the priority 1 schedule is processed until it is empty, proceeded by the priority 2 schedule, proceeded by the priority 3 schedule, etc., until the scheduling queues are empty or the scheduling interval's symbol and/or time budget expires. This process is repeated for each scheduling interval.

[0034] The inter-path scheduler 52 determines fairness utilizing the attributes

(parameters) of the TMC, and with other real time indicators. Inter-path capacities are scheduled with delay sensitivity by considering capacity allocations within the dominant domains of the CP COS (i.e., priorities) as depicted in Figure 6. However within this delay sensitive prioritization, inter-path scheduling accomplishes asymmetrical fairness, by utilizing an arrival time domain windowing scheme instead of a pure FIFO (first in first out) time stamp approach. This scheme treats all capacity requests occurring within the specified scheduling interval, called the arrival time domain window (ATTDW), as equally arriving at the same instance in time regardless of direction. In addition to windowed time stamping, the amount of capacity in each direction is also considered and used as part of the queuing algorithm's weighting. Thus, when priority classification and arrival time are equivalent, capacity grants slightly favor the predictive delay and asymmetry inherent within transmission queue capacities which are future predictors of capacity utilization. This increases scheduling efficiency by reacting to asymmetrical indications as early as possible.

[0035] Other factors are also considered within inter-path scheduling to ensure fairness results within a mix of delay sensitive data streams. Such factors include the CP's asymmetric minimum and peak PDU capacity rates. For example, if one CP peak PDU rate is 2 times greater than another CP peak rate, fairness dictates that this be considered within the context of inter- path scheduling. Furthermore, each CP COS also has a maximum delay increment parameter within its TMC. All other things being equal (e.g., priority-classification, arrival time, queue capacity, and peak rates) the CP with the lower maximum delay increment attribute is favorably weighted. Due to these various weighting parameters that deal with priority, delay sensitivity, and fairness the inter-path scheduler 52 is referred to as utilizing windowed fair priority queuing (WFPQ).

[0036] The windowed fair priority queuing (WFPQ) is used to schedule PDU transmissions at the inter-path level. It determines capacity allocations across the multiple communication paths (CPs) and it provides fair scheduling within a priority-based scheme. In a pure fair-share scheduling scheme, share based capacity allocations for lower priority high capacity remotes may disrupt the delay limits of higher priority PDU streams. This may occur within a fair share scheduling scheme because the path with the largest share of capacity also receives the lowest delay. A purely fixed priority scheduling scheme may be able to accommodate the delay bounds of the high priority PDU streams, but congestion within the higher priority PDU streams may result in periods of capacity starvation for lower priority PDU streams. Windowed fair priority queuing provides a combination of a fair share scheme with a priority based scheme.

[0037] The windowed fair priority queuing (WFPQ) technique is used to assign scheduling priority within the context of multiple CPs across multiple classes-of-service. In such a mode of operation, during times of congestion, some delay may be introduced within the higher priority data streams to allow fair share allocations to the other classes-of-service. The WFPQ technique as presented herein, is utilized independently for each class-of-service at the inter-path level. Thus, the WFPQ technique is used to assign scheduling priority within the context of each class-of-service independently.

[0038] Inter-path fairness within windowed fair-priority queuing as described herein utilized a time domain window (TDW), wherein capacity requests are given an equivalent arrival time classification (AT_TDW)- The arrival time is a function of the time domain window. For example, within a TDD or FDD framing structure using the framing interval as the time domain window, AT_TDW is incremented for each framing interval. All other parameters being equal (e.g., same COS, peak rate, etc.), PDUs arriving earlier in time are prioritized and thus scheduled before PDUs arriving later in time. By giving both forward and reverse capacity requests within a single airlink frame the same arrival time stamps, asymmetric scheduling fairness is maintained.

[0039] The communication path's current class-of-service queue capacity is also utilized in the scheduling prioritization process. This is because queue sizes are finite resources and all other things being equal (e.g., same COS, AT_TD , etc.) the path with queues being stressed receive appropriate priority. Furthermore, indications of the upcoming direction of asymmetry may be predicted from the class-of-service queue capacities, which may be utilized to provide efficient our scheduling.

[0040] The scheduling priority indicator is a function of priority within the time domain window (P_TDW) in accordance with the following equation;

PTDW = Hj + ATTDW + 1 PDUEsuClas. (2) where, P_TDW is the scheduling priority indicator as a function of the time domain window, TDW is the time domain window, μ_j is the priority of the j^,h class-of-service, AT_TDW is the arrival time classification as a function of the time domain window, and 1/ PDU_Est_ci_ass is the reciprocal of the current estimated capacity of the particular class-of-service queue within the applicable communication path. The smaller the resultant values of P_TDW the higher the priority.

[0041] Within a given time domain window, forward and reverse capacity requests receive corresponding treatment and scheduling priority as a function of the arrival time, class- of-service, and queue capacities. As indicated by equation (2), as capacity builds in the resultant reverse link direction, asymmetry is weighted in the reverse link direction. This mechanism provides a forward look ahead tracking efficiency within asymmetrical capacity allocations.

[0042] During times of congestion, PDUs may accumulate in the airlink's distributed queuing mechanism resulting in arriving capacity requests that may have significant numbers of PDU capacities being requested. If the individual PDUs of capacity are not prioritized and scheduled in a distributed fashion proportional to their estimated arrival rate, large numbers (clumps) of lower rate capacities may distort the delay and peak rate requirements of higher rate communication paths. For example, a first communication path may have a peak PDU rate of five (5) PDUs per airlink frame, a second communication path may have a peak PDU rate of ten (10) PDUs per airlink frame, and each communication path request ten (10) PDUs within the same time domain window (TDW) and class-of-service. Furthermore, some packet only remote stations may be serviced within multiple airlink framing rates related to the communication path's TMC minimum frame service rate (MFSR) parameter. Additionally, each communication path has maximum delay increments (MDI) specified per class-of-service . These parameters are real numbers that are functions of the TDW. Thus, the final resultant PDU scheduling priority indicator is also a function of the communication path's peak PDU rate and maximum delay parameters.

[0043] A single PDU schedule is queued for each PDU of capacity being requested (e.g., if a reverse link capacity request asks for 10 newly arrived asynchronous PDUs then 10 PDU schedules are queued). A PDU schedule contains a scheduling priority indicator (P_PD_U) that is equivalent to the real number defined by the following equation.

PPDU = μj + ATTDW + 1 PDU_Est_cia_SS + (AT_Adj - 1 )/R_path-peak + Max-Delay_Ciass, (3) where, μj, ATTDW, 1/ PDU_Est_ciass, and are as previously defined, and AT_Adj is the per PDU arrival time adjustment, Max-Delayciass is the communication path's maximum allowable delay increment for this particular class-of-service (e.g., P2-MDI), and R_path-_peak is the peak rate of PDUs per airlink frame for the appropriate direction (e.g., R_path-p_eak = FPPR/MFSR).

[0044] AT_Adj is the per PDU arrival time adjustment. It is used to distribute scheduling priority among the individual PDUs as a function of the path's peak PDU rate which is directly proportional to its worst case arrival rate. Thus PDU capacity is scheduled at or better than the resultant arrival rate. The delay between their arrival, scheduling, and transmission, is in part, a function of inherent airlink delay, over subscription, and the current congestion within the wireless transmission system. Moreover, this distribution of schedule at peak rate between different communication paths results in a round robin-like attribute of fairness being inherently part of the hub's inter-path PDU scheduling.

[0045] For each PDU within the capacity request, AT_Adj increments from one (1) to the total number of new PDUs requested. Each instance of ATAdj is divided by the path's peak PDU rate to arrive at the specific PDUs arrival time adjustment. For example, if the path peak rate is 10 PDUs per airlink frame and 10 PDUs of capacity were requested, the resultant arrival time adjustments are calculated as follows: 0, 0. 1 , 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, and 0.9, respectively. Such a distribution is scheduled within the same time domain window (TDW). (the adjustment would not result in a value that would effectively increment AT_TDW)-

[0046] However if the path's peak rate was 5 PDUs per airlink frame and 10 PDUs of capacity were requested, the resultant arrival time adjustments are as follows: 0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, 1.4, and 1.8, respectively. Such a capacity request (in excess of peak PDU rate) is another indicator of congestion and in this particular example results in a PDU schedule distribution at path peak rate across two (2) airlink frames, (the adjustment would result in a value that would effectively increment AT_TDW)- This adjustment take into consideration and compensates for remote stations having a minimum frame service rate (MFSR) that is a multiple of the airlink frames (e.g., half rate and quarter rate remotes), and adjusts their scheduling accordingly.

[0047] The Max-Del ayciass parameter is an inter-path scheduling adjustment increment that is provisional within each communication path's TMC. It is a unique parameter for each class-of-service. Thus, within the communication path, each class-of-service may have its own max delay increment (e.g., PI - MDI, P2 - MDI, etc.). Its use within the windowed fair priority queuing (WFPQ) technique provides provisional scheduling differentiation between competing communication paths. Thus, all other things being equal, the communication path with the lower maximum delay value, for the particular class-of-service, will be scheduled ahead of a communication path with a higher max delay tolerance, for the same class-of-service.

[0048] A set of PDU schedules is built and added to the appropriate class-of-service schedule queues every scheduling interval or time domain window. The fair priority queues are also be processed by the hub station's inter-path scheduler 52 every scheduling interval in P_PDU priority order. During a TDD airlink frame TDW, the forward link transmission schedulers update their new PDU bids first and then the reverse link transmission schedulers update their new PDU bids. However since the PDU schedule queues are processed by the inter-path scheduler in P_PDU order, the PDU schedule queues represent a prioritized schedule that inherently defines the next TDD frame's asymmetrical symbol budget attributes.

[0049] Figure 7 is a flow diagram of an exemplary process for scheduling and allocating bandwidth. PDUs are formed at step 70 PDUs, as described previously, herein. Information to be communicated is segmented into fixed sized PDUs. Each PDU comprises a payload portion, and a prepended portion. The payload portion contains data indicative of the information to be transmitted and/or received. The prepended portion contains data indicative of the priority of the information and the amount of time a PDU has been waiting to be scheduled. The inter-path scheduler provides PDU bulk grants to each communication path at step 72, in accordance with an aggregate demand for resources and bandwidth by all remote stations 22, as described previously herein. These PDU bulk grants are received by the intra-path scheduler at step 74, and PDUs are scheduled for transmission, for each communication path at step 76, in accordance with the description herein pertaining to equations (1) and (2) and associated text. At step 78, capacity counts for each communication path/class-of-service are updated and provided to the inter-path scheduler for subsequent scheduling.

[0050] The present invention may be embodied in the form of computer-implemented processes and apparatus for practicing those processes. The present invention may also be embodied in the form of computer program code embodied in tangible media, such as floppy diskettes, read only memories (ROMs), CD-ROMs, hard drives, high density disk, or any other computer-readable storage medium, wherein, when the computer program code is loaded into and executed by a computer, the computer becomes an apparatus for practicing the invention. The present invention may also be embodied in the form of computer program code, for example, whether stored in a storage medium, loaded into and/or executed by a computer, or transmitted over some transmission medium, such as over electrical wiring or cabling, through fiber optics, or via electromagnetic radiation, wherein, when the computer program code is loaded into and executed by a computer, the computer becomes an apparatus for practicing the invention. When implemented on a general-purpose processor, the computer program code segments configure the processor to create specific logic circuits.

[0051] Although the present invention is described in terms of exemplary embodiments, it is not limited thereto. Rather, the appended claims should be construed broadly, to include other variants and embodiments of the invention, which may be made by those skilled in the art without departing from the scope and range of equivalents of the invention.

Claims

Claims

What is claimed is:

1. A method for allocating bandwidth for a communications path of a point to multipoint (PTMP) system comprising a plurality of communications paths, said method comprising the step of allocating available bandwidth within each communications path in accordance with a predetermined priority and an arrival time of information to be transmitted within said available bandwidth.

2. A method in accordance with claim 1 , wherein available bandwidth for each communications path is allocated further in accordance with respective selected attributes of each communications path.

3. A method in accordance with claim 2, further comprising the step of dynamically reconfiguring said selected attributes of each communications path in accordance with an aggregate demand for available bandwidth for all communications paths.

4. A method in accordance with claim 1, further comprising the step of formatting said information to be transmitted into protocol data units (PDUs), wherein:

each PDU comprises a respective header portion, a respective payload portion, and a respective prepended portion; and

each prepended portion comprises a respective priority corresponding to each PDU's respective payload portion and a respective time stamp indicative of an amount of time a respective PDU has been waiting to be allocated.

5. A method in accordance with claim 4, further comprising the steps of:

between two PDUs having different priorities, allocating bandwidth to a PDU having a higher priority, if said PDU having a higher priority has a time stamp less than or equal to a predetermined value; and allocating bandwidth to a PDU having a lower.priority, if said PDU having a lower priority has a time stamp greater than said predetermined value.

6. A method in accordance with claim 4, further comprising the step of dynamically calculating an instantaneous priority for each PDU in accordance with the following equation:

P_j(t) = MDI_j - Cθ_j(t), wherein

t represents time;

P_j(t) represents an instantaneous priority for a j_th class-of-service as a function of time;

MDI_j represents a communication path's maximum allowable delay increment for a j_th class-of-service; and

Cθ_j(t) represents a value indicative of the longest amount of time a PDU of a j_th class-of-service has been waiting for allocation.

7. A method in accordance with claim 4, further comprising the step of dynamically calculating a scheduling priority indicator for each PDU as a function of a time domain window in accordance with the following equation:

PTDW = μj + ATTDW + I/PDU_Est_ciass, wherein

P_TDW represents said scheduling priority indicator;

TDW represents said time domain window;

μ_j represents a priority of a j^th class-of-service;

AT_TDW represents an arrival time classification as a function of a time domain window; and

I/PDUE_St_ciass is a reciprocal of a current estimated capacity of a particular class- of-service queue within an applicable communication path.

8. A method in accordance with claim 4, wherein said attributes comprise at least one of the attributes selected from the group consisting of minimum frame service rate, forward peak PDU rate, reverse peak PDU rate, forward minimum PDU rate, reverse minimum PDU rate, and at least one priority maximum delay increment.

9. A computer readable medium having embodied thereon a program for causing a processor to allocate bandwidth in a point to multipoint (PTMP) system comprising a plurality of communications paths, said computer readable medium comprising means for causing said processor to allocate available bandwidth within each communications path in accordance with a predetermined priority and an arrival time of information to be transmitted within said available bandwidth.

10. A computer readable medium in accordance with claim 9, wherein available bandwidth for each communications path is allocated further in accordance with respective selected attributes of each communications path.

11. A computer readable medium in accordance with claim 10, further comprising means for causing said processor to dynamically reconfigure said selected attributes of each communications path in accordance with an aggregate demand for available bandwidth for all communications paths.

12. A computer readable medium in accordance with claim 9, further comprising means for causing said processor to format said information to be transmitted into protocol data units (PDUs), wherein:

each PDU comprises a respective header portion, a respective payload portion, and a respective prepended portion; and

each prepended portion comprises a respective priority corresponding to each PDU's respective payload portion and a respective time stamp indicative of an amount of time a respective PDU has been waiting to be allocated.

13. A computer readable medium in accordance with claim 12, further comprising: between two PDUs having different priorities, means for causing said processor to allocate bandwidth to a PDU having a higher priority, if said PDU having a higher priority has a time stamp less than or equal to a predetermined value; and

means for causing said processor to allocate bandwidth to a PDU having a lower priority, if said PDU having a lower priority has a time stamp greater than said predetermined value.

14. A computer readable medium in accordance with claim 12, further comprising means for causing said processor to dynamically calculate an instantaneous priority for each PDU in accordance with the following equation:

P_j(t) = MDI_j - Cϋ_j(t), wherein

t represents time;

P_j(t) represents an instantaneous priority for a j_th class-of-service as a function of time;

MDI_j represents a communication path's maximum allowable delay increment for a jn, class-of-service; and

Cύ_j(t) represents a value indicative of the longest amount of time a PDU of a j_th class-of- service has been waiting for allocation.

15. A computer readable medium in accordance with claim 12, further comprising means for causing said processor to dynamically calculate a scheduling priority indicator for each PDU as a function of a time domain window in accordance with the following equation:

PTDW = μj + AT_TDw + I/PDU_Est_ciass, wherein

P_TDW represents said scheduling priority indicator;

TDW represents said time domain window;

μ, represents a priority of a j^th class-of-service;

AT_TDW represents an arrival time classification as a function of a time domain window; and I/PDUEst_ciass is a reciprocal of a current estimated capacity of a particular class- of-service queue within an applicable communication path.

16. A computer readable medium in accordance with claim 12, wherein said attributes comprise at least one of the attributes selected from the group consisting of minimum frame service rate, forward peak PDU rate, reverse peak PDU rate, forward minimum PDU rate, reverse minimum PDU rate, and at least one priority maximum delay increment.