US20140082203A1  Method and apparatus for capacity dimensioning in a communication network  Google Patents
Method and apparatus for capacity dimensioning in a communication network Download PDFInfo
 Publication number
 US20140082203A1 US20140082203A1 US14/089,624 US201314089624A US2014082203A1 US 20140082203 A1 US20140082203 A1 US 20140082203A1 US 201314089624 A US201314089624 A US 201314089624A US 2014082203 A1 US2014082203 A1 US 2014082203A1
 Authority
 US
 United States
 Prior art keywords
 data
 sub
 elastic
 fraction
 priority
 Prior art date
 Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
 Granted
Links
 238000004891 communication Methods 0 claims description title 3
 239000002609 media Substances 0 abstract claims description 11
 239000000727 fractions Substances 0 claims description 31
 239000011159 matrix materials Substances 0 claims description 18
 230000001965 increased Effects 0 claims description 12
 238000000034 methods Methods 0 description 11
 230000015654 memory Effects 0 description 8
 238000004422 calculation algorithm Methods 0 description 5
 230000014509 gene expression Effects 0 description 5
 238000004364 calculation methods Methods 0 description 3
 230000000875 corresponding Effects 0 description 2
 230000001419 dependent Effects 0 description 2
 238000009826 distribution Methods 0 description 2
 238000003860 storage Methods 0 description 2
 230000003044 adaptive Effects 0 description 1
 239000003570 air Substances 0 description 1
 230000006399 behavior Effects 0 description 1
 230000003542 behavioural Effects 0 description 1
 230000001413 cellular Effects 0 description 1
 238000007600 charging Methods 0 description 1
 239000011162 core materials Substances 0 description 1
 230000003247 decreasing Effects 0 description 1
 230000036012 kel Effects 0 description 1
 230000000670 limiting Effects 0 description 1
 238000002620 method outputs Methods 0 description 1
 239000000203 mixtures Substances 0 description 1
 238000010295 mobile communication Methods 0 description 1
 230000003287 optical Effects 0 description 1
 230000002633 protecting Effects 0 description 1
 230000002829 reduced Effects 0 description 1
 239000003826 tablets Substances 0 description 1
 230000001702 transmitter Effects 0 description 1
Images
Classifications

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L65/00—Network arrangements or protocols for realtime communications
 H04L65/80—QoS aspects

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L43/00—Arrangements for monitoring or testing packet switching networks
 H04L43/08—Monitoring based on specific metrics
 H04L43/0876—Network utilization
 H04L43/0882—Utilization of link capacity

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L65/00—Network arrangements or protocols for realtime communications
 H04L65/40—Services or applications
 H04L65/4069—Services related to one way streaming
 H04L65/4084—Content on demand

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L65/00—Network arrangements or protocols for realtime communications
 H04L65/60—Media handling, encoding, streaming or conversion

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L67/00—Networkspecific arrangements or communication protocols supporting networked applications
 H04L67/02—Networkspecific arrangements or communication protocols supporting networked applications involving the use of webbased technology, e.g. hyper text transfer protocol [HTTP]

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L67/00—Networkspecific arrangements or communication protocols supporting networked applications
 H04L67/30—Networkspecific arrangements or communication protocols supporting networked applications involving profiles
 H04L67/306—User profiles

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04W—WIRELESS COMMUNICATION NETWORKS
 H04W28/00—Network traffic or resource management
 H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
 H04W28/24—Negotiating SLA [Service Level Agreement]; Negotiating QoS [Quality of Service]

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04W—WIRELESS COMMUNICATION NETWORKS
 H04W28/00—Network traffic or resource management
 H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
 H04W28/18—Negotiating wireless communication parameters
 H04W28/20—Negotiating bandwidth
Abstract
A method, computer readable medium and apparatus for calculating a capacity for high speed packet access data in a link in a communications network are disclosed. For example, the method initializes parameters associated with streaming data, long elastic data and short elastic data, determines, via a processor, a capacity value such that a quality of service metric is met for the streaming data, the long elastic data and the short elastic data and provisions the link with the capacity value if the quality of service metric is met.
Description
 This application is a continuation of U.S. patent application Ser. No. 12/963,369, filed Dec. 8, 2010, which is currently allowed and is herein incorporated by reference in its entirely.
 Packet based video is one of the fastest growing internet applications. At a high level, video streaming services, as well as any other similar streaming service with realtime neardeterministic and timely delivery requirements, employing Internet protocol (IP) can be classified into two categories: video over IP networks and video over the Internet. In the case of video over IP networks, video is delivered using dedicated IP networks. The availability of dedicated network allows the actuation of various admission control and traffic prioritization policies at the routers and switches to provide higher priority to video and other realtime traffic and, thereby protecting the more stringent quality of service (QoS) requirements of video.
 Video over the Internet, on the other hand, uses public Internet infrastructures to deliver streaming video. This is also referred to as videooverthetop and is growing rapidly. However, video over the Internet is carried just like any other data that is carried via hypertext transfer protocol (HTTP) using transmission control protocol (TCP). As a result, an insufficient amount of bandwidth on various links may be available, thereby causing the QoS requirements of the video to not be met.
 In one embodiment, the present disclosure teaches a method, computer readable medium and apparatus for calculating a capacity for high speed packet access data in a link in a communications network. In one embodiment, the method comprises initializing parameters associated with streaming data, long elastic data and short elastic data, determining, via a processor, a capacity value such that a quality of service metric is met for the streaming data, the long elastic data and the short elastic data and provisioning the link with the capacity value if the quality of service metric is met
 The teaching of the present disclosure can be readily understood by considering the following detailed description in conjunction with the accompanying drawings, in which:

FIG. 1 illustrates one example of a communications network architecture; 
FIG. 2 illustrates a block diagram of a data architecture within a link of the communications network; 
FIG. 3 illustrates a high level flowchart of one embodiment of a method for calculating a capacity for high speed packet access data in a link in a communications network; 
FIG. 4 illustrates another high level flowchart of one embodiment of a method for calculating a capacity for high speed packet access data in a link in a communications network; 
FIG. 5 illustrates a more detailed flowchart of one embodiment of a method for calculating a capacity for high speed packet access data in a link in a communications network; 
FIG. 6 illustrates a more detailed flowchart of a queuing model block; 
FIG. 7 illustrates a more detailed flowchart of a γ matrix computation; 
FIG. 8 illustrates a more detailed flowchart of a main loop; 
FIG. 9 illustrates a more detailed flowchart of a performance estimation block; and 
FIG. 10 illustrates a highlevel block diagram of a generalpurpose computer suitable for use in performing the functions described herein.  To facilitate understanding, identical reference numerals have been used, where possible, to designate identical elements that are common to the figures.
 The present disclosure broadly discloses a method, computer readable medium and an apparatus for calculating a capacity for high speed packet access data in a link in a communications network. In one embodiment, the link of particular interest may be an access link that connects a wireless base station (e.g., in 3G or 4G wireless networks) to a radio access network (RAN), e.g., an IuB link in 3G or an S1u link in LTE.
FIG. 1 is a block diagram depicting one example of a communications network architecture 100 related to the current disclosure. In one embodiment, the communications network architecture 100 comprises a 3G or a 4G cellular network such as a universal mobile telecommunications system (UMTS) network, a long term evolution (LTE) network and the like. However, it should be noted that the communications network architecture 100 may include other types of communications networks such as general packet radio services (GPRS) networks, global system for mobile communication (GSM) networks or enhanced data rates for GSM evolution (EDGE) networks, and the like, by substituting the appropriate hardware and/or hardware configurations for the respective networks. 
FIG. 1 is a block diagram depicting one example of a communications network architecture 100 related to the current disclosure. In one embodiment, the communications network architecture 100 includes one or more user endpoints or user equipment (UE) 102, a base station 104, a radio network controller (RNC) 106, the Internet 108 and an application server (AS) 112.  In one embodiment, the user endpoints 102 may be any type of user device, such as for example, a mobile telephone, a smart phone, a messaging device, a tablet computer, a laptop computer, an air card and the like. The user endpoint 102 may communicate wirelessly with elements of the communications network architecture 100.
 In one embodiment, the base station 104 may be an eNodeB. In one embodiment, the RNC 106 provides features such as packet scheduling, radio resource control (RRC) and handover. In one embodiment, the AS 112 may include a processor and memory as described in
FIG. 10 below and may be used to perform the modeling techniques discussed below. In one embodiment, the RNC 106 and the AS 112 may be part of a core network of the communication network architecture 100.  The communications network architecture 100 may also include additional hardware or network components that are not illustrated depending on the type of network. For example, if the communications network is a UMTS network, the communications network architecture 100 may also include a serving general packet radio services (GPRS) support node (SGSN), a gateway GPRS support node (GGSN) and the like. If the communications network is an LTE network, the communications network architecture 100 may also include mobile management entity (MME), a serving gateway (SGW), a home subscriber server (HSS), a policy and charging rules function (PCRF), a packet data network (PDN) gateway (PGW) and the like. In other words,
FIG. 1 only illustrates a simplified communications network architecture 100 and should not be interpreted as a limitation of the present disclosure.  In one embodiment, data flows from the user endpoint device 102 via various links in the communications network architecture 100 to send or receive data to and from various destinations or sources via the Internet 108. For example,
FIG. 1 illustrates a link 110 between the base station 104 and the RNC 106. In one embodiment, the link 110 may be an luB interface of an UMTS communications network or an S1u link in an LTE communications network. However, the link 110 may be any link or interface of any type of communications network.  It is important for service providers today to accurately engineer the links, e.g., the link 110, in the communications network architecture 100 to have sufficient, yet optimal, capacity to handle the traffic flowing through each link. For example, various types of data may have quality of service (QoS) requirements that must be met based upon contractual agreements with the subscriber.

FIG. 2 illustrates a block diagram of a data architecture 200 within the link 110. In one embodiment, the data flowing through the link 110 may be divided into different types of data. A first portion 204 of the link 110 may include conversational data 210 (e.g., voice, call oriented data, admission controlled data, percall bandwidth data and the like) and stream (R99) data 208 (e.g., audio streaming, call oriented data, admission controlled data, percall bandwidth data, and the like). The capacity for the link 110 to handle conversational data 210 and the stream (R99) 208 data may use one type of modeling, such as for example, a KaufmanRoberts dimensional modeling method.  A second portion 202 of the link 110 may include high speed packet access (HSPA) data 206 (or any other broadband data) that uses transmission control protocol (TCP). HSPA data 206 may include, e.g., web browsing data, flow oriented and elastic data, rate adaptable data, finite access bandwidth data and the like. Previously, dimensioning models used for HSPA data 206 (and other similar wideband data) lumped all the data as a single type of data, i.e., elastic data.
 Elastic data may be defined as a data type that is bit conserving (i.e., it sends all of the data), but where the time for transmission may change depending on the available bandwidth. In other words, there is no time conservation for elastic data and the transmission duration can change.
 However, previous dimensioning models for HSPA data 206 that lumped all of the data as a single type of elastic data are no longer accurate for the second portion 202 of the link 110 carrying the HSPA data 206, in its entirety. Currently, streaming data (e.g., video over the Internet) is becoming a substantial portion of the HSPA data traffic that flows through the link 110. As discussed above, HSPA streaming data is not the same as the streaming data such as the conversational data 210 or the stream (R99) data 208 discussed above. In addition, the HSPA streaming data should not be confused with streaming data that is transmitted via dedicated IP networks. Rather, the streaming data that is part of the HSPA data 206 is transmitted via TCP.
 In one embodiment, streaming data may be defined as time conserving, but not bit conserving. For example, if there is congestion, bits may be dropped (or coding rate reduced at the server) and a downgraded quality of the streaming data may be transmitted. In other words, the duration of the streaming data does not change. For example, if a video is five minutes long, the duration of the transmission will be five minutes.
 In addition to streaming video, the present disclosure also pertains to an elastic data traffic class (i.e., the nonstream portion of the HSPA data 206) that competes against streaming video for bandwidth at various network resources. At a finer granularity, the elastic data sessions are in turn comprised of long elastic sessions (e.g., large file transfer sessions and the like) and short elastic sessions (e.g., email sessions, web surfing sessions and the like) that employ TCP. In the remainder of this disclosure, the former may be abbreviated as long elastic data and the latter as short elastic data.
 Notably, streaming data, long elastic data and short elastic data have unique behavioral attributes and different performance criteria. Therefore, the previously used dimensioning techniques to calculate a required capacity for the link 110 may no longer be accurate.
 In one embodiment, the present disclosure provides a new dimensioning technique to calculate a required capacity for the second portion 202 of the link 110 that takes into consideration the different types of HSPA data 206 (e.g., streaming data, long elastic data and short elastic data). In one embodiment, the new dimensioning technique may be applied as part of a process that can be performed by the AS 112 or a general purpose computer as disclosed in
FIG. 10 . 
FIG. 3 illustrates a high level flowchart of a method 300 for one embodiment of calculating a capacity for HSPA data in a link in a communications network. In one embodiment, the method 300 may be implemented by the AS 112 or a general purpose computer having a processor, a memory and input/output devices as discussed below with reference toFIG. 10 .  The method 300 begins at step 302 and proceeds to step 304. At step 304, the method 300 initializes parameters associated with streaming data, long elastic data and short elastic data. The specific parameters are discussed below with reference to
FIGS. 49 .  At step 306, the method 300 determines, via a processor, a capacity value such that a QoS metric is met for the streaming data, the long elastic data and the short elastic data. In one embodiment, the determining step may be an iterative process. For example, one of the initial parameters may be an initial capacity value. Based upon the initial capacity value, it may be determined if the QoS metric is met for the streaming data, the long elastic data and the short elastic data. If not, the initial capacity value may be revised higher and the determining step may be repeated until the QoS metric is met.
 At step 308, the method 300 provisions the link with the capacity value if the quality of service metric is met. In other words, the capacity value is the appropriate allocation that meets the quality of service metrics. The method 300 ends at step 310.

FIG. 4 illustrates a high level flowchart of a method 400 of another embodiment for calculating a capacity for high speed packet access data in a link in a communications network. In one embodiment, the method 400 may be implemented by the AS 112 or a general purpose computer having a processor, a memory and input/output devices as discussed below with reference toFIG. 10 .  The method 400 begins at step 402 where the method 400 is initialized with various input parameters. In one embodiment, the input parameters may include an aggregate busy hour data traffic intensity for the streaming data (K_{s}), an aggregate busy hour data traffic intensity for the long elastic data (K_{el}), an aggregate busy hour data traffic intensity for the short elastic data (K_{es}), a peak rate of the streaming data (Q_{s}), a packet loss probability for the elastic data (e.g., the long elastic data and the short elastic data) (PLR), a round trip delay time for the elastic data (e.g., the long elastic data and the short elastic data) (RTT), a maximum transmission control protocol segment size (MSS) and a representative file size for the short elastic data (FS).
 In one embodiment, the Q_{s }is dictated by the application drain rate of the streaming data. In other words, Q_{s }is a known fixed value based upon documented attributes of the streaming applications.
 At step 404, the method 400 enters the input parameters from step 402 into the new multiclass capacity dimensioning algorithm. At step 406, the method 400 determines a capacity value (C) that is used to engineer the capacity of a link, e.g., the link 110. In one embodiment, the method 400 may be executed based upon one or more different bandwidth sharing profiles 408 among contending applications. Thus, depending on the sharing profile that is used, the computed outcome for C may vary between profiles.
 In one embodiment, two different types of sharing profiles may be used. A first profile applies an even splitting of capacity among active sessions. For example, in a two dimensional model of x_{s }stream sessions and x_{e }elastic sessions, there may be no distinction among the long and short elastic sessions. In one embodiment, the first profile may be represented as shown below in Equation (1):

$\begin{array}{cc}{\gamma}_{s}\ue8a0\left({x}_{s},{x}_{e}\right)={\gamma}_{e}\ue8a0\left({x}_{s},{x}_{e}\right)=\frac{1}{{x}_{s}+{x}_{e}},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(1\right)\end{array}$  where γ_{s }represents a share of the capacity assigned to each of the x_{s }streaming data sessions and γ_{e }represents a share of the capacity assigned to each of the x_{e }elastic data sessions.
 For a scenario with more than two classes (e.g., a 3class scenario comprised of streaming data, long elastic data and short elastic data), the share of bandwidth captured by each active session belonging to class i, γ_{i}(x), is given by Equation (2):

$\begin{array}{cc}{\gamma}_{i}\ue8a0\left(x\right)=\mathrm{Min}\ue89e\left\{{\alpha}_{i},\frac{1\sum _{j=1}^{i1}\ue89e{x}_{j}\ue89e{\gamma}_{j}\ue8a0\left(x\right)}{\sum _{j=i}^{N}\ue89e{x}_{j}}\right\},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(2\right)\end{array}$  with the convention that the result of any summation where the upper index is less than the lower index equals zero.
 A second profile applies a peak rate proportional bandwidth sharing profile. For example in a two dimensional model of x_{s }stream sessions and x_{e }elastic sessions, the second profile may be represented as shown below in Equations (3) and (4):

$\begin{array}{cc}{\gamma}_{s}\ue8a0\left({x}_{s},{x}_{e}\right)=\frac{{\alpha}_{s}}{{\alpha}_{s}\ue89e{x}_{s}+{\alpha}_{e}\ue89e{x}_{e}},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(3\right)\\ {\gamma}_{e}\ue8a0\left({x}_{s},{x}_{e}\right)=\frac{{\alpha}_{e}}{{\alpha}_{s}\ue89e{x}_{s}+{\alpha}_{e}\ue89e{x}_{e}},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(4\right)\end{array}$  where α_{s }represents a normalized peak rate to pipe capacity for the streaming data and α_{e }represents a normalized peak rate to pipe capacity for the elastic data.
 The α_{s }and the α_{e }may be calculated by Equations (5) and (6):

$\begin{array}{cc}{\alpha}_{s}=\frac{{Q}_{s}}{C},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(5\right)\\ {\alpha}_{e}=\frac{{Q}_{e}}{C},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(6\right)\end{array}$  where Q_{s }represents the peak rate of the streaming data, Q_{e }represents the peak rate of the elastic data and C represents the target capacity of the link, which is being evaluated for QoS compliance in the current instance of the dimensioning process. The value of Q_{s }is determined based upon the type of streaming data, as discussed above.
 The value of Q_{e }may be calculated depending on the type of elastic data. For example, a peak rate of the long elastic data Q_{el }and a peak rate of the short elastic data Q_{es }may be calculated. The values for Q_{el }and Q_{es }may be calculated by considering the behavior of the long elastic and short elastic data under different PLR and RTT. In one embodiment, Equations (7) and (8) below can be used:

$\begin{array}{cc}{Q}_{\mathrm{el}}=\frac{1.22\xb7\mathrm{MSS}}{\mathrm{RTT}\xb7\sqrt{\mathrm{PLR}}},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(7\right)\\ {Q}_{\mathrm{es}}=\frac{\mathrm{FS}}{\mathrm{RTT}\xb7{\mathrm{log}}_{2}\ue8a0\left(\mathrm{FS}/\mathrm{MSS}\right)},& \mathrm{Eq}.\phantom{\rule{0.6em}{0.6ex}}\ue89e\left(8\right)\end{array}$  where the variables, FS, MSS, RTT and PLR are known input parameters, as discussed above.
 As a result, using the Q_{el }and the Q_{es}, the α_{e }may be further categorized into the normalized peak rate for the long elastic data α_{el }and the normalized peak rate for the short elastic data α_{es}. The values for α_{el }and α_{es }are provided below by Equations (9) and (10):

$\begin{array}{cc}{\alpha}_{\mathrm{el}}=\frac{{Q}_{\mathrm{el}}}{C},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(9\right)\\ {\alpha}_{\mathrm{ex}}=\frac{{Q}_{\mathrm{es}}}{C},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(10\right)\end{array}$  For a scenario with more than two classes (e.g., a 3class scenario comprised of streaming data, long elastic data and short elastic data), the share of bandwidth captured by each active session belonging to class i, γ_{i}(x), is given by Equation (11):

$\begin{array}{cc}{\gamma}_{i}\ue8a0\left(x\right)=\frac{{\alpha}_{i}}{\mathrm{Max}\ue89e\left\{1,\sum _{j=1}^{N}\ue89e{\alpha}_{j}\ue89e{x}_{j}\right\}},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(11\right)\end{array}$ 
FIG. 5 illustrates a more detailed flowchart of one embodiment of a method 500 for calculating a capacity for high speed multiclass access packet data in a link in a communications network.FIG. 5 provides more detail to some of the steps illustrated inFIG. 4 . In one embodiment, the method 500 may be implemented by the AS 112 or a general purpose computer having a processor, a memory and input/output devices as discussed below with reference toFIG. 10 .  The method 500 begins at step 502 and proceeds to step 504. At step 504 an initial lower bound on the capacity, C, is set and included in the input parameter list 516. In addition, at step 504, the method 500 sets a variable QoSmet=false.
 The method 500 then proceeds to step 506, where the input parameters from block 516 and a bandwidth sharing profile from block 518 are fed into a queuing model block 506. In one embodiment, the input parameters may include, C, K_{s}, K_{el}, K_{es}, Q_{s}, PLR, RTT, MSS and FS. In one embodiment, the profile may be one of a plurality of different types of sharing profiles. For example, two possible sharing profiles may be an equal sharing profile or a proportional sharing profile. The equations for the two types of sharing profiles are provided by Equations (1)(4) and (11) above.
 At a high level, the queuing model block 506 determines the computational priority ordering of the streaming data, the long elastic data and the short elastic data, and sets the bounds of a three dimensional state transition diagram. It should be noted that if the long elastic data and the short elastic data are lumped together as a single elastic data class, the state transition diagram may be simplified into a two dimensional state transition diagram.
 Attempting to solve the steady state probabilities of the three dimensional state transition diagram and then attempting to estimate from the probabilities performance metrics applicable to each type of data traffic (e.g., the streaming data, the long elastic data and the short elastic data) can be a daunting task. For example, solving the steady state probabilities of the three dimensional state transition diagram would require solving a system of linear equations in K^{3 }variables, having a complexity of approximately K^{6}.
 As a result, in one embodiment the queuing model block 506 attempts to perform conservative bounding of the state transition diagrams for various performance metrics, for example, subpar customer time fraction (SCTF) for the streaming data and a subpar customer data fraction (SCDF) for the elastic data. In one embodiment, the SCDF may be further divided into a long elastic data subpar customer data fraction (SCDF_{el}) and a short elastic data subpar customer data fraction (SCDF_{es}). Additionally, the SCTF can be mapped to an objective streaming video quality metric, such as for example, the probability of stall or the Peak Signal to Noise Ratio (PSNR). These performance metrics are discussed in further detail below.
 For simplicity, the illustrative examples and equations below are provided for a twodimensional state transition diagram. However, the equations may easily be modified for the threedimensional state transition diagram (or any arbitrary Ndimensional state transition diagram). For example, expressions with “e” may be substituted with “(el+es)”. The conservative bounding may be determined by calculating a pessimistic weight function {hacek over (Ψ)}(x_{s},x_{e}) for each state (x_{s},x_{e}). The technique involves following a socalled heaviest weight path along the edges of the state space (e.g., a square for 2D and a cube for 3D). Some of the theoretical underpinnings for performing this technique may be found in a document entitled “On Performance Bounds for the Integration of Elastic and Adaptive Streaming Flows,” T. Bonald and A. Proutiere, SIGMETRICS/Performance' 04, June 2004, New York, which is incorporated by reference in its entirety. Alternatively, an optimistic weight function {circumflex over (Ψ)}(x_{s},x_{e}) for each state (x_{s},x_{e}) that follows the lightest weight path along the edges of the state space may also be used to arrive at optimistic estimates for the performance metrics.
 In one embodiment, the pessimistic weight function {hacek over (Ψ)}(x_{s},x_{e}) and the optimistic weight function {circumflex over (Ψ)}(x_{s},x_{e}) may be calculated by first ordering the classes of sessions. For a given state of a system (x_{1}, x_{2}), denote the departure rate of class i sessions (i=1,2) by φ_{i}(x_{1},x_{2}), i=1,2.
 To exemplify the notion of departure rates in a twoclass system comprised of stream (x_{1}) an elastic (x_{2}), if the average duration of each streaming session is denoted by 1/μ_{s }(wherein μ_{s }represents an inverse of the streaming session duration), then the total rate at which stream sessions are departing the system when there are x_{1 }active stream sessions equals φ_{1}(x_{1},x_{2})=x_{1}μ_{s}. Note in particular that φ_{1}(x_{1},x_{2}) is independent of the number of elastic sessions x_{2}. This is reflective of the fact that stream sessions are time conserving as described earlier and last for fixed durations, independent of the congestion levels (while they are nonbit conserving). Now, suppose the average duration of an elastic session, if the entire capacity of the link were to be at its disposal, is denoted by 1/μ_{e }(wherein μ_{e }represents an inverse of the elastic session duration). Then the rate at which elastic sessions are departing the system when there are x_{1 }stream sessions and x_{2 }elastic sessions would be equal to φ_{2}(x_{1},x_{2})=x_{2}μ_{e}×γ_{e}(x_{1}, x_{2}), where γ_{e}(x_{1},x_{2}) is the profiledependent bandwidth share available to each elastic session given by Equations 14 and Equation 11. The fact that φ_{2}(x_{1},x_{2}) depends on both the stream and elastic occupancies (i.e., x_{1 }and x_{2}) ensues from the fact that, unlike stream traffic, elastic traffic is bit conserving but not time conserving. In other words, the duration of an elastic session stretches or shrinks depending on the congestion level, till the last bit is successfully transmitted.
 Apart from the above definitions for the average stream session duration 1/μ_{s }and elastic session duration subject to full capacity availability 1/μ_{e }(1/μ_{el }for long elastic and 1/μ_{es }for short elastic in a 3class model), we also define the mean stream session arrival rate λ_{s }and elastic session arrival rate λ_{e }(λ_{el }for long elastic and λ_{es }for short elastic in a 3class model). In terms of these parameters, we define the normalized stream session Erlangs ρ_{s}=λ_{s}/μ_{s }and the normalized elastic session erlangs ρ_{e}=λ_{e}/μ_{e }(ρ_{el}=λ_{el}/μ_{el }for long elastic and ρ_{es}=λ_{es}/μ_{es }for short elastic in a 3class model). While the ρ parameters are necessary for the computational sequence to be described, estimates for the arrival rates λ and average inverse durations μ are typically not available. However, as shown later, the ρ parameters are alternately calculated in terms of the known aggregate busy hour traffic intensities or volumes {K_{s},K_{el},K_{es}} and peak rates {Q_{s}, Q_{el},Q_{es}}.
 With the concept of departure rates well defined, let the classes be ordered such that:

$\begin{array}{cc}\frac{{\varphi}_{1}\ue8a0\left({x}_{1},{x}_{2}1\right)}{{\varphi}_{1}\ue8a0\left({x}_{1},{x}_{2}\right)}\le \frac{{\varphi}_{2}\ue8a0\left({x}_{1}1,{x}_{2}\right)}{{\varphi}_{2}\ue8a0\left({x}_{1},{x}_{2}\right)},\forall {x}_{1},{x}_{2}\ue89e\text{:}\ue89e{x}_{1}>0,{x}_{2}>0,& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(12\right)\end{array}$  Given the ordered classes, the pessimistic and optimistic weight functions are given by Equations (13) and (14) below:

$\begin{array}{cc}\stackrel{\u22d3}{\Psi}\ue8a0\left({x}_{1},{x}_{2}\right)=\frac{{\lambda}_{1}^{{x}_{1}}\ue89e{\lambda}_{2}^{{x}_{2}}}{\prod _{i=1}^{{x}_{2}}\ue89e{\varphi}_{2}\ue8a0\left({x}_{2},i\right)\ue89e\prod _{i=1}^{{x}_{1}}\ue89e{\varphi}_{1}\ue8a0\left(i,0\right)},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(13\right)\\ \stackrel{\u22d2}{\Psi}\ue8a0\left({x}_{1},{x}_{2}\right)=\frac{{\lambda}_{1}^{{x}_{1}}\ue89e{\lambda}_{2}^{{x}_{2}}}{\prod _{i=1}^{{x}_{1}}\ue89e{\varphi}_{2}\ue8a0\left(i,{x}_{2}\right)\ue89e\prod _{i=1}^{{x}_{2}}\ue89e{\varphi}_{1}\ue8a0\left(0,1\right)},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(14\right)\end{array}$  wherein λ's, as defined earlier, represent mean session arrival rates and {hacek over (Ψ)}(0,0)={circumflex over (Ψ)}(0,0)=1. As discussed above, the φ terms in the denominators of the above equations contain μ terms, which divide into corresponding λ terms in the numerator such that these equations can be restated in terms of terms of ρ terms. While λ's and μ's are typically unavailable, the ρ terms can be calculated alternately from traffic volumes and peak rates as shown later, thus facilitating the computational steps to be described.
 Over the remainder of this description, the infinite statespace {(0,0) . . . (∞,∞)} is approximated by a finite statespace {(0,0) . . . (K,K)}, with K being chosen so as to achieve an acceptable degree of numerical accuracy. Similarly, in the context of a 3dimensional model, the infinite statespace {(0,0,0) . . . (∞,∞,∞)} will be approximated by the finite statespace {(0,0,0) . . . (K, K, K)}.
 Note that Equation (13) may be expressed in the following efficient recursive format:

$\begin{array}{cc}\stackrel{\u22d3}{\Psi}\ue8a0\left(0,0\right)=1;\ue89e\text{}\ue89e\stackrel{\u22d3}{\Psi}\ue8a0\left({x}_{1},0\right)=\stackrel{\u22d3}{\Psi}\ue8a0\left({x}_{1}1,0\right)\times \frac{{\lambda}_{1}}{{\varphi}_{1}\ue8a0\left({x}_{1},0\right)},{x}_{1}=1,\dots \ue89e\phantom{\rule{0.8em}{0.8ex}},K\ue89e\text{}\ue89e\stackrel{\u22d3}{\Psi}\ue8a0\left({x}_{1},{x}_{2}\right)=\stackrel{\u22d3}{\Psi}\ue8a0\left({x}_{1},{x}_{2}1\right)\times \frac{{\lambda}_{2}}{{\varphi}_{1}\ue8a0\left({x}_{1},{x}_{2}\right)},{x}_{2}=1,\dots \ue89e\phantom{\rule{0.8em}{0.8ex}},K& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(13\ue89eA\right)\end{array}$  Based on the recursive variant Equation (13A), the following algorithmic sequence may be followed to exhaustively compute the weight functions {{hacek over (Ψ)}(0,0) . . . {hacek over (Ψ)}(K,K)} for all the system states {[0,0] . . . [K, K]}:

Set {hacek over (Ψ)}(0,0) = 1; For (i=0 to K){ If (i>0) then calculate {hacek over (Ψ)}(0,i) from Equation (13A); For (j=1 to K) Calculate {hacek over (Ψ)}(j,i) from Equation (13A); }  As may be appreciated, Equation (14) for the optimistic bound can also be expressed in an analogous recursive format with the associated efficient implementation. The recursive formats and associated efficient implementations may be used in the algorithmic descriptions to be given.
 In one embodiment, given the pessimistic weight function above in Equation (13), the pessimistic steady state distribution, {hacek over (π)}, for each state, x, may be computed by Equation (15) below:

{hacek over (π)}(x _{s} ,x _{e})={hacek over (π)}(0,0){hacek over (Ψ)}(x _{s} ,x _{e}), where {hacek over (π)}(0,0)=1/Σ_{(x} _{ s } _{,x} _{ e } _{)}{hacek over (Ψ)}(x _{s} ,x _{e}), Eq. (15)  Using the calculated pessimistic bounds, the conservative (pessimistic) bounds of the performance metrics SCTF and SCDF can be calculated using Equations (16) and (17) below:

$\begin{array}{cc}\mathrm{SCTF}=\frac{\sum _{\left({x}_{s},{x}_{e}\right)\ue89e\text{:}\ue89e{\gamma}_{s}\ue8a0\left({x}_{s},{x}_{e}\right)<\mathrm{SST}}\ue89e{x}_{s}\ue89e\pi \ue8a0\left({x}_{s},{x}_{e}\right)}{\sum _{\left({x}_{s},{x}_{e}\right)}\ue89e{x}_{s}\ue89e\pi \ue8a0\left({x}_{s},{x}_{e}\right)},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(16\right)\\ \mathrm{SCDF}=\frac{\sum _{\left({x}_{s},{x}_{e}\right)\ue89e\text{:}\ue89e{\gamma}_{e}\ue8a0\left({x}_{s},{x}_{e}\right)<\mathrm{ET}}\ue89e{x}_{s}\ue89e{\gamma}_{e}\ue8a0\left({x}_{s},{x}_{e}\right)\ue89e\pi \ue8a0\left({x}_{s},{x}_{e}\right)}{\sum _{\left({x}_{s},{x}_{e}\right)}\ue89e{x}_{s}\ue89e{\gamma}_{e}\ue8a0\left({x}_{s},{x}_{e}\right)\ue89e\pi \ue8a0\left({x}_{s},{x}_{e}\right)},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(17\right)\end{array}$  where ET represents a customer stipulated data throughput rate acceptability threshold for each elastic session (e.g., 800 kilobytes per second (Kbps)), SST represents a customer stipulated data throughput rate acceptability threshold for each stream session (e.g., 800 Kbps) and γ_{e}(x_{s},x_{e}) is defined by Equation (4) above. In the 3dimensional model, it should be noted that the ET would have two distinct components, ELT and EST, for long elastic data throughput rate acceptability threshold and short elastic data throughput rate acceptability threshold, respectively, where appropriate.
 It should be noted that any of the equations above for two classes may be generalized to an arbitrary number of classes. For example, any of the equations may be generalized for three classes to be applied specifically to a system having the streaming data, the long elastic data and the short elastic data as distinct components.
 The method 500 then proceeds to a performance estimation block 508. Based upon the bounds calculated above, the performance metrics such as SCTF, SCDF_{el }and SCDF_{es}, may be calculated via equations, such as Equations (16) and (17) above.
 The method 500 then proceeds to step 510, where the performance metrics are compared against a target value for each of the performance metrics. For example, there may be a predefined SCTF target value (TGT_{SCTF}), a predefined long elastic SCDF target value (TGT_{SCDF} ^{long}) and a predefined short elastic SCDF target value (TGT_{SCDF} ^{short}). In one embodiment, the TGT_{SCTF}, TGT_{scDF} ^{long }and TGT_{SCDF} ^{short }may be configurable values. For example, the TGT_{SCTF }may be set to 0.1%, the TGT_{SCDF} ^{long }may be set to 10% and the TGT_{SCDF} ^{short }may be set to 1%. The numerical values are only provided as examples and should not be considered limiting. If the each one of the performance metrics, SCTF, SCDF_{el }and SCDF_{es}, calculated in step 508 are less than or equal to their respective target values, TGT_{SCTF}, TGT_{SCDF} ^{long }and TGT_{SCDF} ^{short}, then the variable QoSmet is set to “true”.
 At step 512, the method 500 determines whether the QoSmet variable is true or false. If the variable QoSmet is false, then the method 500 proceeds to step 514, where the value of C is updated in the input parameter list 516. In other words, at step 514 the method 500 determined that the initial value of C was not a sufficient amount of capacity for the QoS to be met for each of the different types of traffic depending on the type of sharing profile that was used to perform the calculations. As a result, another iteration of the calculations must be performed with the intent of converging to the optimal value for C.
 In one embodiment, the search for optimum capacity, C, is effectuated by linearly increasing the value of C, as shown in
FIG. 5 . Alternatively, a faster logarithmic search method may be used to facilitate the determination for the optimum capacity, C. To illustrate, if the initial value of C was set as 10,000 Kbps, and the initial value of C determined to be insufficient, the value of C may be exponentially increased to 20,000 Kbps, then to 40,000 kbps, and then to 80,000 Kbps and so forth, until a smallest upper bound (e.g., 40,000 Kbps) that satisfies the metrics, and a highest lower bound (e.g., 20,000 Kbps) that does not satisfy the metrics, are identified.  With the bounds available, the performance evaluation algorithm is run for a capacity, C, equal to the midpoint between the upper and lower bounds (i.e., 30,000 Kbps, for the example quoted). If the metrics are satisfied at the midpoint, then the midpoint becomes a new upper bound and the lower bound is maintained at its previous value. Otherwise the midpoint becomes a new lower bound and the upper bound is maintained at its previous value. Now the procedure is repeated using the new pair of lower and upper bounds. The ensuing recursion is continued till the upper and lower bounds are within an acceptable margin of error, at which point the value of the upper bound is output as the optimal capacity C. As may be appreciated, this version of the search outer loop of the algorithm implementation involves a logarithmic number of steps, and would converge significantly faster than a linear implementation, shown for illustrative purposes in
FIG. 5 . Further details for performing the logarithmic search may be found with reference to copending U.S. patent application Ser. No. 12/655,236 filed on Dec. 23, 2009, which is incorporated by reference in its entirety.  From step 514, the method resubmits the input parameters with the updated value of C back into the queuing model block 506 and the method is repeated until QoSmet=true at step 512. At step 512, if the variable QoSmet=true, then the method 500 proceeds to step 520. At step 520, the method outputs C as the engineered capacity and the method 500 ends.

FIG. 6 illustrates a more detailed flowchart of a method 600 that is one embodiment of a way to perform the queuing model block 506 illustrated inFIG. 5 . It should be noted that the description of the algorithm captured in method 600 shown inFIG. 6 , and further refined in methods 700900 shown inFIGS. 79 below, approximate the infinite state space by a finite state space given by {[0,0,0], . . . ,[K,K,K]}, where K is chosen so as to achieve a desired level of numerical accuracy (e.g., K=100).  The method 600 starts at step 602 and proceeds to 604. At step 604, the method 600 receives the input parameters from block 516 as illustrated in
FIG. 5 . The input parameters are used to calculate various values, such as for example, Q_{el}, Q_{es}, ρ_{s}, ρ_{el}, ρ_{es}, α_{s}, α_{el }and α_{es}. The equations for calculating the values of Q_{el}, Q_{es}, α_{s}, α_{el }and α_{es }are discussed above in Equations (5) and (7)(10). The symbols ρ_{s}, ρ_{el }and ρ_{es}, as defined earlier, represent a normalized stream session erlangs, a normalized long elastic session erlangs and a normalized short elastic session erlangs. The values for ρ_{s}, ρ_{el }and ρ_{es }may be alternatively calculated using the Equations (18)(20) below: 
$\begin{array}{cc}{\rho}_{s}=\frac{{K}_{s}}{{Q}_{s}},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(18\right)\\ {\rho}_{\mathrm{el}}=\frac{{K}_{\mathrm{el}}}{C},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(19\right)\\ {\rho}_{\mathrm{es}}=\frac{{K}_{\mathrm{es}}}{C},& \mathrm{Eq}.\phantom{\rule{0.8em}{0.8ex}}\ue89e\left(20\right)\end{array}$  Once the various normalized values are calculated, the method 600 proceeds to step 606 where a γ matrix computation is performed. As discussed above, γ represents a capacity assigned to a particular session. The γ matrix provides a matrix that describes the persession capacity assigned to each of the particular types of sessions γ_{i}(x), under the exhaustive set of states of the system {x} (the system state being defined by a vector of the number of active sessions belonging to the distinct classes under consideration). It should be noted that the values of the γ matrix entries would depend on the particular bandwidth sharing profile specified.
 The method 600 then proceeds to step 608, wherein the highest algorithmic priority, index value of 1, is assigned to the streaming data, s. It should be noted that this assignment does not imply any kind of traffic prioritization in the physical world. For example, absolute trafficdependent priorities are feasible only in dedicated private networks, and not in the HTTP/TCP public internet scenario. Rather, the above assignment has meaning only in an algorithmic sense, and is a consequence of the ordering implied by Equation (12), in light of the property described above, whereby the departure rates of stream sessions are independent of congestion and hence independent of the number of active elastic sessions, while the departure rates of elastic sessions (long, as well as short) do depend on congestion, hence on the occupancy levels of all classes.
 At step 610, the method 600 then determines which bandwidth sharing profile was applied. For example, referring back to
FIG. 5 , at step 518, a bandwidth sharing profile was selected and entered into the queuing model block 506. If the equal bandwidth sharing profile was selected, then the method 600 proceeds to step 612. At step 612, the method 600 maps indices 2 and 3 to the long elastic data and the short elastic data according to the mathematical expression “Map indices 2, 3ε{el, es}: α_{2}≦α_{3}.” In other words, under the equal bandwidth sharing profile, the indices for the two elastic subclasses are assigned in the order of increasing peak rates, which is the appropriate ordering to satisfy Equation (12) in this context.  However, if the proportional bandwidth sharing profile was selected, then the method 600 proceeds to step 614. At step 614, the method 600 maps indices 2 and 3 to the long elastic data and the short elastic data according to the mathematical expression “Map indices 2, 3ε{el,es}:α_{2}≧α_{3}.” In other words, under the proportional bandwidth sharing profile, the indices for the two elastic subclasses are assigned in the order of decreasing peak rates, which is the appropriate ordering to satisfy Equation (12) in this context.
 At step 616, the method 600 initializes each state variable x_{i}, for i=13, to 0 and the initial weight associated with state [0 0 0], π[0,0,0], to 1. Mathematically, this is expressed as shown in step 616 and below in Equation (21):

x _{1} =x _{2} =x _{3}=0;π[0,0,0]=norm=1; Eq. (21)  The method 600 then proceeds to step 618 where the main loop is performed. In one embodiment, the main loop calculates the steady state probabilities for each state within the bounds of the pessimistic weight function or optimistic weight function, as discussed earlier. The details of the main loop are described further below with reference to
FIG. 8 , in the context of conservative bounds; as will be explained, the sequence of certain operations therein may be reversed to arrive at optimistic bounds instead.  At step 620, the method 600 remaps the steady state distribution, π, in accordance with the original ordering of the streaming data, the long elastic data and the short elastic data. The method 600 then proceeds to step 622, where the method 600 exits at step 622 back to the performance estimation block 508, illustrated in
FIG. 5 . 
FIG. 7 illustrates a more detailed flowchart of a method 700 for the γ matrix computation 606 illustrated inFIG. 6 . In one embodiment, the γ matrix provides information as to how bandwidth is being shared by the active session belonging to competing data types (e.g., stream session data sessions, long elastic data sessions and short elastic data sessions) for each state of the state transition diagram. The method 700 starts at step 702 and then proceeds to step 704. At step 704, the method 700 indexes the three traffic types in order of increasing peak rates as shown, and initializes the values for the γ matrix. The computation begins at state {0,0,0} for x_{1}, x_{2 }and x_{3}. Mathematically, it is shown in step 704 and below in Equation (22): 
Map indices 1,2,3ε{s,el,es}:α _{1}≦α_{2}≦α_{3} ;x _{1} =x _{2} =x _{3}=0; Eq. (22)  The method 700 then proceeds to step 706, where the method 700 determines which bandwidth sharing profile was selected. Again, referring back to
FIG. 5 , a bandwidth sharing profile was selected in step 518. Depending on the bandwidth sharing profile selected, the method 700 proceeds to step 708 or to step 710. If the equal bandwidth sharing profile was selected, the method 700 at step 708 calculates the γ matrix according to Equation (2) provided above for γ_{1}{x_{1},x_{2},x_{3}}, γ_{2}{x_{1},x_{2},x_{3}} and γ_{3}{x_{1},x_{2},x_{3}}. However, if the proportional bandwidth sharing profile was selected, the method 700 at step 710 calculates the γ matrix according to Equation (11) above for i=1, 2, 3.  The method 700 then proceeds through each state of the state transition diagram until the bandwidth share for each state is calculated in the γ matrix as shown by steps 712, 714, 716, 718 and 720. Again, as mentioned earlier, K, is the maximum occupancy level of each class considered (i.e., the approximation of the infinite statespace by a finite one). In other words, steps 712720 represent mathematically steps for ensuring that all the states are visited. In this case, for example, in the order [0,0,0], [1,0,0], . . . [K,0,0], [0,1,0], . . . [0,0,K], . . . [K,K,K].
 Once the γ matrix is computed, the method 700 proceeds to step 722 where the bandwidth shares for each index are remapped to the traffic classes. For example, the γ_{1 }may be mapped to γ_{s}, the γ_{2 }may be mapped to γ_{el }and the γ_{3 }may be mapped to γ_{es }(if α_{1}≦α_{2}≦α_{3}). The method 700 ends at 724 and returns to step 606 of
FIG. 6 . 
FIG. 8 illustrates a more detailed flowchart of a method 800 for performing the main loop step 618 ofFIG. 6 .FIG. 8 illustrates one embodiment that applies the pessimistic weight function, {hacek over (Ψ)}, discussed above. With that objective, method 800 goes through the heaviest weight path starting with x_{3}, then x_{2 }and x_{1}, as shown. It should be noted that if the optimistic weight function, {circumflex over (Ψ)}, were to be used then the method 800 may be modified to go through the lightest weight path starting with x_{1 }to x_{3}. The probability variables π[0,0,0] . . . π[K,K,K] are initially used to compute and store the weight functions {hacek over (Ω)}[0,0,0] . . . [K,K,K]; subsequently, a normalization step is applied to convert them to probabilities. In particular π[0,0,0]={hacek over (Ψ)}[0,0,0] was initialized to 1 in step 616 prior to the invocation of method 800.  The method 800 starts at step 802 and then proceeds to step 804. At step 804, the method 800 sets x_{3}=1 and starts calculating the steady state probabilities via the heaviest weight path. The method 800 proceeds to step 806.
 Step 806 along with the loop test in step 808 executes Equation (13A) in conjunction with the associated efficient implementation, for the highest index 3. In particular, {hacek over (Ψ)}[0,0,1] through {hacek over (Ψ)}[0,0,K] are computed in the first pass.
 Next, x_{2 }is incremented to 1 in step 810, determines if x_{1 }is less than K at step 812 and the method 800 proceeds to step 816 where ti[0,1,0] is computed via the logic of Equation (13A).
 Next, x_{3 }is reset to 1 in step 804 and the algorithm returns to step 806 to compute {hacek over (Ψ)}[0,1,1] through {hacek over (Ψ)}[0,1,K]. This circuit involving steps 806, 808, 810, 812, 818 and 804 continues until {hacek over (Ψ)}[0,0,1] through {hacek over (Ψ)}[0,K,K] are generated.
 At this point, the method 800 proceeds to step 814 and increments x_{1 }to 1. At step 816 if the test passes (i.e., x_{1}<K is yes), the method 800 proceeds to step 820 where {hacek over (Ψ)}[1,0,0] is computed. The distinctions between the departure rates (φ's) for stream and elastic as described above, may be noted by comparing the calculation in step 820 (for stream) against that in step 806 or 818 (for elastic classes). The method 800 proceeds to step 822 where the variable norm is computed and x_{2 }is reset to 0. It should be noted that the variable norm may be continually updated upon computation of the weight for each and every state in steps 806, 818, 822 and 822.
 Next the method 800 proceeds to step 804 where x_{3 }is reset to 1, following which {hacek over (Ψ)}[1,0,1] is computed in step 806, The circuit involving steps 806, 808, 810, 812, 818 and 804 is invoked a second time until {hacek over (Ψ)}[1,0,1] through {hacek over (Ψ)}[1,K, K] are generated.
 At this point, method 800 again proceeds to step 814 to increment x_{1}. It may thus be appreciated that the macro circuit involving steps 806, 808, 810, 812, 814, 816, 818, 804, 820 and 822 eventually generates all of {hacek over (Ψ)}[0,0,1] through {hacek over (Ψ)}[K,K,K], at which point the test fails in step 816 (i.e. x_{1}<K is no) bringing method 800 to step 824. As may be noticed, the value of the normalization variable norm at this point equals the sum of the weight function values for all the system states (the infinite statespace being approximated as K^{3 }finite states). In step 824, the weight function value for each state is divided by norm to convert to the corresponding state probability. The method 800 then proceeds to step 826 where the method 800 ends and returns to step 618 of
FIG. 6 .  It is again noted that in a different embodiment, the optimistic probability bounds may be computed instead of the pessimistic bounds simply by reversing the order of the indices. For example, by using x_{1 }in place of x_{3 }in steps 804, 806 and 808 and x_{3 }in the place of x_{1 }in steps 814, 816 and 820, the computation sequence can be changed to generate the optimistic weight functions {{circumflex over (Ψ)}[0,0,0] . . . {circumflex over (Ψ)}[K,K,K]} instead of the pessimistic weight functions as is being accomplished in
FIG. 8 with the order shown. 
FIG. 9 illustrates a more detailed flowchart of a method 900 for performing the performance estimation block step 508 ofFIG. 5 . The method 900 starts at step 902 from the queuing model block 506 ofFIG. 5 and proceeds to step 904. At step 904, the method 900 initializes values to begin calculation of the performance metric parameters SCTF, SCDF_{el }and SCDF_{es}. For example, x_{1}, x_{2 }and x_{3 }are set to 0. In addition, the parameters SCTFdenom, SCTF numer, SCDFdenom[el], SCDFnumer[el], SCDFdenom[es] and SCDFnumer[el] are also set to 0. The parameters SCTF numer, SCDFnumer[el] and SCDFnumer[es] represent the numerator value (e.g., the numerator of Equations (16) and (17)) for calculating the performance metric parameters SCTF, SCDF_{el }and SCDF_{es}, respectively. The parameters SCTFdenom, SCDFdenom[el] and SCDFdenom[es] represent the denominator value (e.g., the denominator of Equations (16) and (17)) for calculating the performance metric parameters SCTF, SCDF_{el }and SCDF_{es}, respectively.  The method 900 then proceeds to step 906. The method 900 at step 906 calculates the parameters SCTFdenom, SCTF numer, SCDFdenom[el], SCDFnumer[el], SCDFdenom[es] and SCDFnumer[es] for each state. For example, the calculation is defined by Equations (16) and (17) described above. Step 906 illustrates another way of expressing what is shown in Equations (16) and (17) described above. It should be noted that the mathematical expression “A+=B” is shorthand representation for the mathematical expression “A=A+B.”
 The method 900 proceeds through each state as shown by steps 908, 910, 912, 914 and 916 until the final values of the parameters SCTFdenom, SCTF numer, SCDFdenom[el], SCDFnumer[el], SCDFdenom[es] and SCDFnumer[es] are calculated. In other words, steps 908916 represent mathematically steps for ensuring that all the states are visited. In this case, for example, in the order [0,0,0], [1,0,0], . . . [K,0,0], [0,1,0], . . . [0,0,K], . . . [K,K,K].
 At step 918, the values of the performance metric parameters SCTF, SCDF_{el }and SCDF_{es }are calculated using the final values of the parameters SCTFdenom, SCTF numer, SCDFdenom[el], SCDFnumer[el], SCDFdenom[es] and SCDFnumer[es]. The method 900 proceeds to step 920 where the values of SCTF, SCDF_{el }and SCDF_{es }are outputted to step 510 of
FIG. 5 .  It should be noted that although not explicitly specified, one or more steps of the methods described herein may include a storing, displaying and/or outputting step as required for a particular application. In other words, any data, records, fields, and/or intermediate results discussed in the methods can be stored, displayed, and/or outputted to another device as required for a particular application. Furthermore, steps or blocks in
FIGS. 3 and 9 that recite a determining operation, or involve a decision, do not necessarily require that both branches of the determining operation be practiced. In other words, one of the branches of the determining operation can be deemed as an optional step. 
FIG. 10 depicts a highlevel block diagram of a generalpurpose computer suitable for use in performing the functions described herein. As depicted inFIG. 10 , the system 1000 comprises a processor element 1002 (e.g., a CPU), a memory 1004, e.g., random access memory (RAM) and/or read only memory (ROM), a module 1005 for calculating a capacity for high speed packet access data in a link in a communications network, and various input/output devices 1006 (e.g., storage devices, including but not limited to, a tape drive, a floppy drive, a hard disk drive or a compact disk drive, a receiver, a transmitter, a speaker, a display, a speech synthesizer, an output port, and a user input device (such as a keyboard, a keypad, a mouse, and the like)).  It should be noted that the present disclosure can be implemented in software and/or in a combination of software and hardware, e.g., using application specific integrated circuits (ASIC), a general purpose computer or any other hardware equivalents. In one embodiment, the present module or process 1005 for calculating a capacity for high speed packet access data in a link in a communications network can be loaded into memory 1004 and executed by processor 1002 to implement the functions as discussed above. As such, the present method 1005 for calculating a capacity for high speed packet access data in a link in a communications network (including associated data structures) of the present disclosure can be stored on a nontransitory (tangible or physical) computer readable storage medium, e.g., RAM memory, magnetic or optical drive or diskette and the like.
 While various embodiments have been described above, it should be understood that they have been presented by way of example only, and not limitation. Thus, the breadth and scope of a preferred embodiment should not be limited by any of the abovedescribed exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.
Claims (20)
1. A method for calculating a capacity for high speed packet access data in a link in a communications network, comprising:
determining, by a processor, a capacity value where a quality of service metric is met for streaming data, long elastic data and short elastic data, wherein the determining comprises:
determining an ordering of priority for the streaming data, the long elastic data and the short elastic data;
calculating a subpar customer time fraction for the streaming data, a long elastic subpar customer data fraction and a short elastic subpar customer data fraction based upon the ordering of priority; and
determining if the subpar customer time fraction is greater than or equal to a target subpar customer time fraction, if the long elastic subpar customer data fraction is greater than or equal to a target long elastic subpar customer data fraction, and if the short elastic subpar customer data fraction is greater than or equal to a target short elastic subpar customer data fraction; and
provisioning, by the processor, the link with the capacity value if the quality of service metric is met.
2. The method of claim 1 , wherein the streaming data, the long elastic data and the short elastic data employ a transmission control protocol.
3. The method of claim 1 , wherein parameters associated with the streaming data, the long elastic data and the short elastic data, comprise an initial capacity value, an aggregate busy hour data traffic intensity for the streaming data, the long elastic data and the short elastic data, a peak rate of the streaming data, a packet loss probability for the long elastic data and the short elastic data, a round trip delay time for the long elastic data and the short elastic data, a maximum transmission control protocol segment size and a representative file size for the short elastic data.
4. The method of claim 1 , wherein the determining the ordering of priority comprises:
calculating a bandwidth share matrix based upon a bandwidth sharing profile;
initializing an ordering of priority to provide an initial ordering of priority;
calculating a bound based upon the bandwidth share matrix and the initial ordering of priority based upon the bandwidth sharing profile; and
remapping the order of priority based upon the bound.
5. The method of claim 4 , wherein the bandwidth sharing profile comprises an equal bandwidth sharing profile or a peak rate proportional bandwidth sharing profile.
6. The method of claim 4 , wherein the bound comprises a pessimistic bound that follows a heaviest weight path.
7. The method of claim 4 , wherein the bound comprises an optimistic bound that follows a lightest weight path.
8. The method of claim 4 , wherein the streaming data is mapped as a highest priority.
9. The method of claim 1 , further comprising:
increasing an initial capacity value if the quality of service metric is not met;
determining if the quality of service metric is met with an increased capacity value; and
repeating the increasing and the determining if the quality of service metric is not met with the increased capacity value until the quality of service metric is met and the capacity value is determined.
10. The method of claim 1 , wherein the link is an luB interface of a universal mobile telecommunications system.
11. The method of claim 1 , wherein the link is an S1 interface of a long term evolution communication network.
12. A tangible computerreadable medium storing a plurality of instructions which, when executed by a processor, cause the processor to perform operations for calculating a capacity for high speed packet access data in a link in a communications network, the operations comprising:
determining a capacity value where a quality of service metric is met for streaming data, long elastic data and short elastic data, wherein the determining comprises:
determining an ordering of priority for the streaming data, the long elastic data and the short elastic data;
calculating a subpar customer time fraction for the streaming data, a long elastic subpar customer data fraction and a short elastic subpar customer data fraction based upon the ordering of priority; and
determining if the subpar customer time fraction is greater than or equal to a target subpar customer time fraction, if the long elastic subpar customer data fraction is greater than or equal to a target long elastic subpar customer data fraction, and if the short elastic subpar customer data fraction is greater than or equal to a target short elastic subpar customer data fraction; and
provisioning the link with the capacity value if the quality of service metric is met.
13. The tangible computer readable medium of claim 12 , wherein the determining the ordering of priority comprises:
calculating a bandwidth share matrix based upon a bandwidth sharing profile;
initializing an ordering of priority to provide an initial ordering of priority; and
calculating a bound based upon the bandwidth share matrix and the initial ordering of priority based upon the bandwidth sharing profile;
remapping the order of priority based upon the bound.
14. The tangible computer readable medium of claim 13 , wherein the bandwidth sharing profile comprises an equal bandwidth sharing profile or a peak rate proportional bandwidth sharing profile.
15. The tangible computer readable medium of claim 13 , wherein the bound comprises a pessimistic bound that follows a heaviest weight path.
16. The tangible computer readable medium of claim 13 , wherein the bound comprises an optimistic bound that follows a lightest weight path.
17. The tangible computer readable medium of claim 12 , further comprising:
increasing an initial capacity value if the quality of service metric is not met;
determining if the quality of service metric is met with an increased capacity value; and
repeating the increasing and the determining if the quality of service metric is not met with the increased capacity value until the quality of service metric is met and the capacity value is determined.
18. An apparatus for calculating a capacity for high speed packet access data in a link in a communications network, comprising:
a processor; and
a computer readable medium storing a plurality of instructions which, when executed by the processor, cause the processor to perform operations, the operations comprising:
determining a capacity value where a quality of service metric is met for streaming data, long elastic data and short elastic data, wherein the determining comprises:
determining an ordering of priority for the streaming data, the long elastic data and the short elastic data;
calculating a subpar customer time fraction for the streaming data, a long elastic subpar customer data fraction and a short elastic subpar customer data fraction based upon the ordering of priority; and
determining if the subpar customer time fraction is greater than or equal to a target subpar customer time fraction, if the long elastic subpar customer data fraction is greater than or equal to a target long elastic subpar customer data fraction, and if the short elastic subpar customer data fraction is greater than or equal to a target short elastic subpar customer data fraction; and
provisioning the link with the capacity value if the quality of service metric is met.
19. The apparatus of claim 18 , wherein the determining the ordering of priority comprises:
calculating a bandwidth share matrix based upon a bandwidth sharing profile;
initializing an ordering of priority to provide an initial ordering of priority; and
calculating a bound based upon the bandwidth share matrix and the initial ordering of priority based upon the bandwidth sharing profile;
remapping the order of priority based upon the bound.
20. The apparatus of claim 19 , wherein the bandwidth sharing profile comprises an equal bandwidth sharing profile or a peak rate proportional bandwidth sharing profile.
Priority Applications (2)
Application Number  Priority Date  Filing Date  Title 

US12/963,369 US8595374B2 (en)  20101208  20101208  Method and apparatus for capacity dimensioning in a communication network 
US14/089,624 US9270725B2 (en)  20101208  20131125  Method and apparatus for capacity dimensioning in a communication network 
Applications Claiming Priority (2)
Application Number  Priority Date  Filing Date  Title 

US14/089,624 US9270725B2 (en)  20101208  20131125  Method and apparatus for capacity dimensioning in a communication network 
US15/050,157 US9935994B2 (en)  20101208  20160222  Method and apparatus for capacity dimensioning in a communication network 
Related Parent Applications (1)
Application Number  Title  Priority Date  Filing Date  

US12/963,369 Continuation US8595374B2 (en)  20101208  20101208  Method and apparatus for capacity dimensioning in a communication network 
Related Child Applications (1)
Application Number  Title  Priority Date  Filing Date 

US15/050,157 Continuation US9935994B2 (en)  20101208  20160222  Method and apparatus for capacity dimensioning in a communication network 
Publications (2)
Publication Number  Publication Date 

US20140082203A1 true US20140082203A1 (en)  20140320 
US9270725B2 US9270725B2 (en)  20160223 
Family
ID=46200549
Family Applications (3)
Application Number  Title  Priority Date  Filing Date 

US12/963,369 Expired  Fee Related US8595374B2 (en)  20101208  20101208  Method and apparatus for capacity dimensioning in a communication network 
US14/089,624 Active 20310213 US9270725B2 (en)  20101208  20131125  Method and apparatus for capacity dimensioning in a communication network 
US15/050,157 Active 20310129 US9935994B2 (en)  20101208  20160222  Method and apparatus for capacity dimensioning in a communication network 
Family Applications Before (1)
Application Number  Title  Priority Date  Filing Date 

US12/963,369 Expired  Fee Related US8595374B2 (en)  20101208  20101208  Method and apparatus for capacity dimensioning in a communication network 
Family Applications After (1)
Application Number  Title  Priority Date  Filing Date 

US15/050,157 Active 20310129 US9935994B2 (en)  20101208  20160222  Method and apparatus for capacity dimensioning in a communication network 
Country Status (1)
Country  Link 

US (3)  US8595374B2 (en) 
Families Citing this family (4)
Publication number  Priority date  Publication date  Assignee  Title 

US8595374B2 (en)  20101208  20131126  At&T Intellectual Property I, L.P.  Method and apparatus for capacity dimensioning in a communication network 
EP2915044A4 (en) *  20121105  20160720  Sea Street Technologies Inc  Systems and methods for provisioning and managing an elastic computing infrastructure 
US9860535B2 (en) *  20150520  20180102  Integrated Device Technology, Inc.  Method for timedependent visual quality encoding for broadcast services 
US10454989B2 (en) *  20160219  20191022  Verizon Patent And Licensing Inc.  Application quality of experience evaluator for enhancing subjective quality of experience 
Citations (39)
Publication number  Priority date  Publication date  Assignee  Title 

US6636497B1 (en) *  19981130  20031021  Nokia Networks Oy  Air interface capacity scheduling method 
US20040136379A1 (en) *  20010313  20040715  Liao Raymond R  Method and apparatus for allocation of resources 
US20050259616A1 (en) *  20040519  20051124  Tamas Major  Data transport in UMTS 
US20050259623A1 (en) *  20040513  20051124  Harinath Garudadri  Delivery of information over a communication channel 
US20060268788A1 (en) *  20050525  20061130  Motorola, Inc.  Method and apparatus for improved channel maintenance signaling 
US20060276195A1 (en) *  20050603  20061207  Nikus Nordling  Wireless highspeed data network planning tool 
US20070171830A1 (en) *  20060126  20070726  Nokia Corporation  Apparatus, method and computer program product providing radio network controller internal dynamic HSDPA flow control using one of fixed or calculated scaling factors 
US20070265017A1 (en) *  20051130  20071115  Ntt Docomo, Inc.  Call admission control device and call admission control method 
US20070275760A1 (en) *  20040505  20071129  Telefonaktiebolaget Lm Ericsson (Publ)  Hdspa Flow Control, Control Frames Rtt Measurement 
US20080002582A1 (en) *  20060629  20080103  Jinhui Li  CreditBased Wireless Network Scheduling 
US20080019305A1 (en) *  20060720  20080124  Armin Dekorsy  Allocating channels in multiuser or multiservice realtime transmissions of wireless packet data 
US20080125133A1 (en) *  20061125  20080529  Mazen Chmaytelli  System and method for adaptable multimedia download resulting in efficient airlink usage 
US20080186862A1 (en) *  20070205  20080807  Telefonaktiebolaget Lm Ericsson (Publ)  Congestion/load indication for high speed packet access 
US20080212525A1 (en) *  20070302  20080904  Janne Tervonen  Using device profile to determine the most suitable resource reservation for an application 
US20080240045A1 (en) *  20070330  20081002  Gaviphat Lekutai  Method and apparatus for allocation of radio resources 
US20080259812A1 (en) *  20050916  20081023  Telefonaktiebolaget Lm Ericsson (Publ)  Dimensioning Methods For Hsdpa Traffic 
US20090028103A1 (en) *  20050704  20090129  Shanghai Ultimate Power Communications Technology Co., Ltd.  MultiCarrier Hsdpa Channel Establishing Method and the MultiCarrier Downlink Packet Data Transmitting Method 
US20090132691A1 (en) *  20071115  20090521  Societe Francaise De Radiotelephone  Method and system to manage communications 
US20090161688A1 (en) *  20071221  20090625  SunUk Park  Methods and apparatus to allocate shared resources of a high speed packet access channel in a communication network 
US20090180430A1 (en) *  20080110  20090716  Fadell Anthony M  Apparatus and methods for network resource allocation 
US20090201823A1 (en) *  20050524  20090813  Telecom Italia S.P.A.  Method for Dimensioning A Data Packets Handler Apparatus In A PacketSwitched Mobile Communications Network 
US20090215462A1 (en) *  20050826  20090827  Peter Lundh  Flow Control In UMTS 
US20100034185A1 (en) *  20041025  20100211  Telefonaktiebolaget Lm Ericsson  Radio quality based channel resource management 
US20100034087A1 (en) *  20060928  20100211  Nokia Siemens Networks S.P.A.  Controlling congestion detection in hsdpa systems 
US20100157895A1 (en) *  20081031  20100624  Interdigital Patent Holdings, Inc.  Method and apparatus for handling uplink transmissions using multiple uplink carriers 
US20100192212A1 (en) *  20090128  20100729  Gregory G. Raleigh  Automated device provisioning and activation 
US20100238803A1 (en) *  20071101  20100923  Telefonaktiebolaget Lm Ericsson (Publ)  Efficient Flow Control in a Radio Network Controller (RNC) 
US20100248643A1 (en) *  20090331  20100930  At&T Intellectual Property I, L.P.  Intelligent adaptive recoding for improved communications resource utilization 
US20100260049A1 (en) *  20071101  20101014  Telefonaktiebolaget L M Ericsson (Publ)  Limiting RLC Window Size in The HSDPA Flow Control 
US20100322224A1 (en) *  20071214  20101223  Electronics And Telecommunications Research Institute  Server, terminal and method for end to end channel capacity estimation in high speed downlink packet access network 
US20110007645A1 (en) *  20071120  20110113  Nicholas Andrew Tompson  System and process for dimensioning a cellular telecommunications network 
US20110019563A1 (en) *  20090612  20110127  Clara Serrano Solsona  Scheduling Traffic in a Communication Channel 
US20110116460A1 (en) *  20091109  20110519  Movik Networks, Inc.  Burst packet scheduler for improved ran efficiency in umts/hspa networks 
US20110134890A1 (en) *  20041203  20110609  Gabor Fodor  Method and apparatus for allocating radio resources in a mobile radio network 
US20110299393A1 (en) *  20100603  20111208  Almutairi Abdulmohsen M  Adaptive CQIbased HSDPA flow control method 
US20120002544A1 (en) *  20100702  20120105  Nec Laboratories America, Inc.  Dynamic Resource Partitioning for LongTerm Fairness to NonElastic Traffic on a Cellular Basestation 
US20120281536A1 (en) *  20090612  20121108  Cygnus Broadband, Inc.  Systems and methods for detection for prioritizing and scheduling packets in a communication network 
US20120327779A1 (en) *  20090612  20121227  Cygnus Broadband, Inc.  Systems and methods for congestion detection for use in prioritizing and scheduling packets in a communication network 
US20130021909A1 (en) *  20100118  20130124  Telefonaktiebolaget L M Ericsson (Publ)  Method for Reserving Capacity on a Communication Network Link 
Family Cites Families (51)
Publication number  Priority date  Publication date  Assignee  Title 

US6266322B1 (en) *  19980212  20010724  At&T Corp.  Dimensioning bandwidth and connection admission control for elastic traffic in highspeed communication networks 
US6795399B1 (en) *  19981124  20040921  Lucent Technologies Inc.  Link capacity computation methods and apparatus for designing IP networks with performance guarantees 
US6788646B1 (en) *  19991014  20040907  Telefonaktiebolaget Lm Ericsson (Publ)  Link capacity sharing for throughputblocking optimality 
AU2536101A (en) *  20000114  20010724  Qariba Ltd  Resource allocation 
DE60042965D1 (en) *  20000524  20091029  Sony Deutschland Gmbh  QoS negotiation 
US7099346B1 (en) *  20010515  20060829  Golden Bridge Technology, Inc.  Channel capacity optimization for packet services 
US6934555B2 (en) *  20010629  20050823  Telefonaktiebolaget Lm Ericsson (Publ)  Software analysis tool for CDMA system 
US20030069963A1 (en) *  20010927  20030410  Nikil Jayant  System and method of quality of service signaling between client and server devices 
US7944839B2 (en) *  20011130  20110517  British Telecommunications Public Limited Company  Method of resource control in a wireless network 
ITTO20021009A1 (en) *  20021120  20040521  Telecom Italia Lab Spa  A method, system and computer program product for the 
US7225267B2 (en) *  20030127  20070529  Microsoft Corporation  Reactive bandwidth control for streaming data 
EP1595368B1 (en) *  20030219  20070103  Fujitsu Limited  Packet scheduling 
WO2004088937A1 (en) *  20030331  20041014  Corvil Limited  A method and system for quality of service optimisation in a data network 
US7734805B2 (en) *  20030415  20100608  AlcatelLucent Usa Inc.  Method for scheduling transmissions in communication systems 
US7394768B2 (en) *  20030430  20080701  Nokia Corporation  Fair scheduling with guaranteed minimum parameter 
GB0323244D0 (en) *  20031003  20031105  Fujitsu Ltd  Uplink scheduling 
JP4546399B2 (en) *  20031127  20100915  ユーティーシダカン（ジョングオ）ヨウシャンゴンシ  Radio channel resource allocation and rate control method in code division multiple access communication system 
US7961650B2 (en) *  20040216  20110614  Christopher Michael Davies  Network architecture 
JP3821823B2 (en) *  20040506  20060913  松下電器産業株式会社  Wireless communication terminal device and wireless communication method 
US7586848B1 (en) *  20040607  20090908  Nortel Networks Limited  Elastic traffic marking for multipriority packet streams in a communications network 
EP1605641A1 (en) *  20040608  20051214  Matsushita Electric Industrial Co., Ltd.  Mapping of shared physical channels depending on the quality of service class 
EP1605642A1 (en) *  20040608  20051214  Matsushita Electric Industrial Co., Ltd.  Service dependent shared physical channel mapping 
MX2007004384A (en) *  20041029  20070605  Ericsson Telefon Ab L M  Resource allocation in communication networks. 
AT554567T (en) *  20041229  20120515  Ericsson Telefon Ab L M  Priority carrier in a mobile telecommunications network 
US20070002750A1 (en) *  20050701  20070104  Nec Laboratories America, Inc.  Generic Real Time Scheduler for Wireless Packet Data Systems 
US8054826B2 (en) *  20050729  20111108  Alcatel Lucent  Controlling service quality of voice over Internet Protocol on a downlink channel in highspeed wireless data networks 
ES2368158T3 (en) *  20050826  20111114  Telefonaktiebolaget L M Ericsson (Publ)  Method and provision for flow control in the umts making use of information in the ubs field. 
US20070070894A1 (en) *  20050926  20070329  Fan Wang  Method to determine a scheduling priority value for a user data connection based on a quality of service requirement 
KR101168212B1 (en) *  20050930  20120725  텔레콤 이탈리아 소시에떼 퍼 아찌오니  Method For Planning A Cellular Mobile Telecommunications Network 
US7564788B2 (en) *  20051202  20090721  Telefonaktiebolaget Lm Ericsson (Publ)  Flow control for low bitrate users on highspeed downlink 
EP1858210A1 (en) *  20060519  20071121  Whitestein Information Technology Group AG  Method and system for adaptive communication service access 
US7656882B2 (en) *  20060628  20100202  Samsung Electronics Co., Ltd.  Method and system for peak scheduling in a wireless network 
US7864802B1 (en) *  20060817  20110104  Sprint Communications Company L.P.  TCPaware resource allocation in wireless networks 
US8954045B2 (en) *  20060929  20150210  Qualcomm Incorporated  Method and apparatus for managing resources at a wireless device 
US8280994B2 (en) *  20061027  20121002  Rockstar Bidco Lp  Method and apparatus for designing, updating and operating a network based on quality of experience 
US8289851B2 (en) *  20070430  20121016  Alcatel Lucent  Lightweight bandwidthmanagement scheme for elastic traffic 
US20090029645A1 (en) *  20070725  20090129  Teenay Wireless, Inc.  MultiTier Backhaul Network System with Traffic Differentiation and Advanced Processing Capabilities and Methods Therefor 
US8014280B2 (en) *  20070928  20110906  Qualcomm Incorporated  Admission control based on QoS performance in a wireless communication network 
CN101494877B (en) *  20080123  20120822  华为技术有限公司  Admittance control method and apparatus of radio communication system 
EP2245800B1 (en) *  20080212  20120208  Telefonaktiebolaget LM Ericsson (publ)  Allocation and priority handling of uplink and downlink resources 
MX2010010332A (en) *  20080326  20101004  Ericsson Telefon Ab L M  Available end to end bandwidth estimation in packetswitched communication networks. 
WO2009124349A1 (en) *  20080407  20091015  Seeker Wireless Pty Limited  Location of wireless mobile terminals 
US7796514B2 (en) *  20081211  20100914  At&T Intellectual Property I, L.P.  System and method for multiservices packet network traffic engineering 
US9414401B2 (en) *  20081215  20160809  At&T Intellectual Property I, L.P.  Opportunistic service management for elastic applications 
JP5388332B2 (en) *  20090107  20140115  株式会社Ｎｔｔドコモ  Base station apparatus and information transmission method 
US8656284B2 (en) *  20090417  20140218  Empirix Inc.  Method for determining a quality of user experience while performing activities in IP networks 
US8537699B2 (en) *  20090616  20130917  Qualcomm Incorporated  Managing video adaptation algorithms 
US8406129B2 (en) *  20091223  20130326  At&T Intellectual Property I, L.P.  Technique for determining transport capacity required to achieve controllable worst case throughput 
EP2612534B1 (en) *  20100903  20190710  Telefonaktiebolaget LM Ericsson (publ)  Scheduling multiple users on a shared communication channel in a wireless communication system 
US9060032B2 (en) *  20101101  20150616  Seven Networks, Inc.  Selective data compression by a distributed traffic management system to reduce mobile data traffic and signaling traffic 
US8595374B2 (en)  20101208  20131126  At&T Intellectual Property I, L.P.  Method and apparatus for capacity dimensioning in a communication network 

2010
 20101208 US US12/963,369 patent/US8595374B2/en not_active Expired  Fee Related

2013
 20131125 US US14/089,624 patent/US9270725B2/en active Active

2016
 20160222 US US15/050,157 patent/US9935994B2/en active Active
Patent Citations (39)
Publication number  Priority date  Publication date  Assignee  Title 

US6636497B1 (en) *  19981130  20031021  Nokia Networks Oy  Air interface capacity scheduling method 
US20040136379A1 (en) *  20010313  20040715  Liao Raymond R  Method and apparatus for allocation of resources 
US20070275760A1 (en) *  20040505  20071129  Telefonaktiebolaget Lm Ericsson (Publ)  Hdspa Flow Control, Control Frames Rtt Measurement 
US20050259623A1 (en) *  20040513  20051124  Harinath Garudadri  Delivery of information over a communication channel 
US20050259616A1 (en) *  20040519  20051124  Tamas Major  Data transport in UMTS 
US20100034185A1 (en) *  20041025  20100211  Telefonaktiebolaget Lm Ericsson  Radio quality based channel resource management 
US20110134890A1 (en) *  20041203  20110609  Gabor Fodor  Method and apparatus for allocating radio resources in a mobile radio network 
US20090201823A1 (en) *  20050524  20090813  Telecom Italia S.P.A.  Method for Dimensioning A Data Packets Handler Apparatus In A PacketSwitched Mobile Communications Network 
US20060268788A1 (en) *  20050525  20061130  Motorola, Inc.  Method and apparatus for improved channel maintenance signaling 
US20060276195A1 (en) *  20050603  20061207  Nikus Nordling  Wireless highspeed data network planning tool 
US20090028103A1 (en) *  20050704  20090129  Shanghai Ultimate Power Communications Technology Co., Ltd.  MultiCarrier Hsdpa Channel Establishing Method and the MultiCarrier Downlink Packet Data Transmitting Method 
US20090215462A1 (en) *  20050826  20090827  Peter Lundh  Flow Control In UMTS 
US20080259812A1 (en) *  20050916  20081023  Telefonaktiebolaget Lm Ericsson (Publ)  Dimensioning Methods For Hsdpa Traffic 
US20070265017A1 (en) *  20051130  20071115  Ntt Docomo, Inc.  Call admission control device and call admission control method 
US20070171830A1 (en) *  20060126  20070726  Nokia Corporation  Apparatus, method and computer program product providing radio network controller internal dynamic HSDPA flow control using one of fixed or calculated scaling factors 
US20080002582A1 (en) *  20060629  20080103  Jinhui Li  CreditBased Wireless Network Scheduling 
US20080019305A1 (en) *  20060720  20080124  Armin Dekorsy  Allocating channels in multiuser or multiservice realtime transmissions of wireless packet data 
US20100034087A1 (en) *  20060928  20100211  Nokia Siemens Networks S.P.A.  Controlling congestion detection in hsdpa systems 
US20080125133A1 (en) *  20061125  20080529  Mazen Chmaytelli  System and method for adaptable multimedia download resulting in efficient airlink usage 
US20080186862A1 (en) *  20070205  20080807  Telefonaktiebolaget Lm Ericsson (Publ)  Congestion/load indication for high speed packet access 
US20080212525A1 (en) *  20070302  20080904  Janne Tervonen  Using device profile to determine the most suitable resource reservation for an application 
US20080240045A1 (en) *  20070330  20081002  Gaviphat Lekutai  Method and apparatus for allocation of radio resources 
US20100260049A1 (en) *  20071101  20101014  Telefonaktiebolaget L M Ericsson (Publ)  Limiting RLC Window Size in The HSDPA Flow Control 
US20100238803A1 (en) *  20071101  20100923  Telefonaktiebolaget Lm Ericsson (Publ)  Efficient Flow Control in a Radio Network Controller (RNC) 
US20090132691A1 (en) *  20071115  20090521  Societe Francaise De Radiotelephone  Method and system to manage communications 
US20110007645A1 (en) *  20071120  20110113  Nicholas Andrew Tompson  System and process for dimensioning a cellular telecommunications network 
US20100322224A1 (en) *  20071214  20101223  Electronics And Telecommunications Research Institute  Server, terminal and method for end to end channel capacity estimation in high speed downlink packet access network 
US20090161688A1 (en) *  20071221  20090625  SunUk Park  Methods and apparatus to allocate shared resources of a high speed packet access channel in a communication network 
US20090180430A1 (en) *  20080110  20090716  Fadell Anthony M  Apparatus and methods for network resource allocation 
US20100157895A1 (en) *  20081031  20100624  Interdigital Patent Holdings, Inc.  Method and apparatus for handling uplink transmissions using multiple uplink carriers 
US20100192212A1 (en) *  20090128  20100729  Gregory G. Raleigh  Automated device provisioning and activation 
US20100248643A1 (en) *  20090331  20100930  At&T Intellectual Property I, L.P.  Intelligent adaptive recoding for improved communications resource utilization 
US20110019563A1 (en) *  20090612  20110127  Clara Serrano Solsona  Scheduling Traffic in a Communication Channel 
US20120281536A1 (en) *  20090612  20121108  Cygnus Broadband, Inc.  Systems and methods for detection for prioritizing and scheduling packets in a communication network 
US20120327779A1 (en) *  20090612  20121227  Cygnus Broadband, Inc.  Systems and methods for congestion detection for use in prioritizing and scheduling packets in a communication network 
US20110116460A1 (en) *  20091109  20110519  Movik Networks, Inc.  Burst packet scheduler for improved ran efficiency in umts/hspa networks 
US20130021909A1 (en) *  20100118  20130124  Telefonaktiebolaget L M Ericsson (Publ)  Method for Reserving Capacity on a Communication Network Link 
US20110299393A1 (en) *  20100603  20111208  Almutairi Abdulmohsen M  Adaptive CQIbased HSDPA flow control method 
US20120002544A1 (en) *  20100702  20120105  Nec Laboratories America, Inc.  Dynamic Resource Partitioning for LongTerm Fairness to NonElastic Traffic on a Cellular Basestation 
NonPatent Citations (2)
Title 

A. Berger, Y. Kogan, Dimensioning bandwidth for elastic traffic in highspeed data networks, IEEE/ACM Trans. on Networking, Vol 8, No 5, October 2000. * 
EuroNGI, Design and Engineering of the Next Generation Internet, towards convergent multiservice networks, Deliverable No: D.WP.JR.2.1.1On the State of the Art in Controlled Bandwidth Sharing, Information Society Technologies, August 21, 2004, Chapter  A survey on statistical bandwidth sharing  page 23 * 
Also Published As
Publication number  Publication date 

US20160173558A1 (en)  20160616 
US9270725B2 (en)  20160223 
US9935994B2 (en)  20180403 
US20120151078A1 (en)  20120614 
US8595374B2 (en)  20131126 
Similar Documents
Publication  Publication Date  Title 

Choi et al.  Adaptive bandwidth reservation and admission control in QoSsensitive cellular networks  
Song et al.  Predictability of WLAN mobility and its effects on bandwidth provisioning  
US7616572B2 (en)  Call admission control/session management based on N source to destination severity levels for IP networks  
US8018925B2 (en)  System and method for multiservices packet network traffic engineering  
US7843832B2 (en)  Dynamic bandwidth allocation apparatus and method  
CN100581118C (en)  Method and system for evaluating number of additional admissible calls for use in call admission control  
RU2316127C2 (en)  Spectrally limited controlling packet transmission for controlling overload and setting up calls in packetbased networks  
US8953443B2 (en)  Method and apparatus for providing congestion management for a wireless communication network  
CN100348071C (en)  Method and system for optimising the performance of a network  
US20050041584A1 (en)  AutoIP traffic optimization in mobile telecommunications systems  
EP1570686B1 (en)  Method of call admission control in a wireless network background  
US8280994B2 (en)  Method and apparatus for designing, updating and operating a network based on quality of experience  
US20050050246A1 (en)  Method of admission control  
JP2012525780A (en)  Adaptive speed control based on overload signal  
US8335161B2 (en)  Systems and methods for network congestion management using radio access network congestion indicators  
US7660244B2 (en)  Method and apparatus for qualityofservice based admission control using a virtual scheduler  
US8780720B2 (en)  Radio access network load and condition aware traffic shaping control  
JP3924536B2 (en)  Evaluation method of network for mobile communication equipment  
US7787492B2 (en)  Resource allocation management  
Tong et al.  Adaptive call admission control under quality of service constraints: a reinforcement learning solution  
KR20140024851A (en)  System to share network bandwidth among competing applications  
EP2832140B1 (en)  Adaptive traffic management in cellular wireless networks  
CN101601249A (en)  Methods and systems for resizing multimedia content based on quality and rate information  
US20040032828A1 (en)  Service management in cellular networks  
CN101415210B (en)  Method and equipment for determining objective network 
Legal Events
Date  Code  Title  Description 

AS  Assignment 
Owner name: AT&T INTELLECTUAL PROPERTY I, L.P., NEVADA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:PUTHENPURA, SARAT;BLACKBURN, STUART;GHOSAL, DIPAK;AND OTHERS;SIGNING DATES FROM 20101202 TO 20110308;REEL/FRAME:031964/0575 

STCF  Information on status: patent grant 
Free format text: PATENTED CASE 

MAFP  Maintenance fee payment 
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 4 