This application claims priority of U.S. Provisional Patent Application Serial No. 61/313,872, filed March 15, 2010, the disclosure of which is incorporated herein by reference in its entirety. This application is a continuation-in-part of U.S. Patent Application Serial No. 12/696,378, filed January 29, 2010, which claims priority of U.S. Provisional Patent Application Serial No. 61/148,454, filed January 30, 2009, the disclosures of which are incorporated herein by reference in their entireties .

Background

One of the most popular, and network intensive, computer applications is video or multimedia playback over the internet. Several years ago, it was believed that the internet, with its aging protocols and lack of guaranteed delivery, would be unable to support video playback at a quality satisfactory to users. However, applications such as YouTube and Netflix have shown that quality video playback is possible.

There were various mechanisms that were created to improve the user experience. For example, in some embodiments, the client, which has a media player, requests and receives, the entire multimedia file before beginning playback of the file. This insures that, once the file begins playing, it will be able to continue uninterrupted. While this may be acceptable for shorter clips, such a scheme is unacceptable for longer files for several reasons. First, the user is forced to wait until the entire file is downloaded before seeing any portion of the clip. Such a wait may be unacceptable. Second, in an environment with limited network bandwidth, the entire file is transmitted, even if the user only watches the first few seconds of the video.

Another method is known as progressive download. In this embodiment, the client begins downloading the multimedia file from the server. When a certain threshold is reached, such as 3 seconds of the download has been completed, the media player on the client begins displaying the video. The threshold used may be fixed, or may be based on the resolution of the file, the average available network bandwidth or other parameters. This has several advantages over the previous method. First, the user does not need to wait until the entire file is downloaded before beginning to view the video. Secondly, bandwidth is potentially saved if the user chooses to navigate away from the video before it is completely downloaded. However, one shortcoming of this mechanism is that once the video has begun playing, the available bandwidth must remain above a minimum level to insure that the downloading of each subsequent portion of the video completes before that portion needs to be displayed. If the bandwidth decreases, the video may appear choppy, or may pause in order to allow the buffer to fill again.

Progressive download methods typically download the multimedia content at the maximum rate supported by the underlying transport within the limits of the player buffer (i.e. limiting the downloaded content not to overflow the player buffer by TCP acknowledgement mechanisms) . These methods improve the quality of experience (QOE) by reducing stutters since the content is transferred to the client buffer, thus minimizing the possibility of player stopping during content presentation. However, one disadvantage is that when the transit network bandwidth is high compared to the media stream rate, significant content is downloaded to the player, and if the user cancels the current presentation, such as by moving on to a different media-clip, the network bandwidth used for the downloaded content that is not viewed is wasted.

Another mechanism used by Adobe Systems, Inc. is known as Real Time Messaging Protocol (RTMP) . This mechanism controls content delivery using RTMP protocol over TCP, by controlled streaming of the content. Thus, the streaming mechanism maintains a window of what is delivered to the player relative to what is currently being played to the user. Thus bandwidth wasted due to user cancellations of active sessions is reduced, by limiting the content delivered to locations relatively close to the current display position. RTMP has also options to use HTTP or HTTPS as transport options.

Another mechanism is Real Time Streaming Protocol/ Real-time Transport Protocol (RTSP/RTP) . In this protocol, the server controls the rate of delivery of the content to the presentation rate, rather than delivering at the maximum rate of the underlying transport, such as TCP or UDP over IP.

In addition, several HTTP streaming protocols have been defined to deliver live or stored multimedia content at controlled rate (determined by the server or based on cooperation between the client and server) that matches the multimedia stream rate.

RTMP and HTTP streaming methods pace the content delivery, by limiting the rate at which content is delivered. RTMP uses protocol components in the client and the server, where as HTTP Streaming uses additional tags in the HTTP Requests and Responses.

In Adaptive Bit Rate Streaming (ABR Streaming) , the client monitors the bandwidth to the server at the start up of multi-media session and at regular intervals during the multi-media play. Based on the monitored bandwidth, the client selects alternative encodings with different screen resolutions of the same content. The switching between alternative resolutions of the same media content is done explicitly by the client. Several applications, such as Ankeena TV, use ABR streaming over HTTP for live content delivery.

However, in this protocol, it is the client that determines the available bandwidth. Thus, in a radio access network, where devices are routinely added and removed from a particular cell, the client's limited knowledge of the overall network utilization may compromise its ability to accurately predict a suitable encoding and screen resolution in a timely manner.

Figure 1 shows a traditional 3G/UMTS network, including an UE (user equipment) 107, NodeB (or base transceiver station) 106, RNC (radio network controller or base station controller) 105, SGSN (Serving GPRS support node) 104, and a GGSN (gateway GPRS Support node) 103. Also shown in Figure 1 is the protocol used to communicate between these various devices. For example, IuB 108 is the protocol used between Node B 106 and the RNC 105. Similarly, IuPS is the protocol used between the RNC 105 and the SGSN 104. Gn 110 is used between the SGSN 104 and the GGSN 103. Finally, Gi 111 is the IP based interface between the GGSN 103 and the internet.

It would be advantageous if a component in the radio access network, with visibility to total network traffic, were able to determine when a particular user was requesting multimedia content and based on that total network traffic, were able to configure the encoding and resolution of that multimedia content to maximize the user's quality of experience. In addition, it would be beneficial if that component could decouple wireless RAN traffic from core network traffic, by buffering multimedia content and delivering this content in a just-in-time manner to the end user device.

Summary

A network device, capable of understanding communications between an end user and the core network on a RAN network is disclosed. In some embodiments, the device is able to decode the control plane and the user plane. As such, it is able to determine when the end user has requested multimedia content. Once this is known, the device can optimize the delivery of that content in several ways. In one embodiment, the device requests the content from the content server (located in the core network) and transmits this content in a just- in-time manner to the end user. In another embodiment, the device automatically changes/limits the options for encoding and resolution of the content available to the client, or pro-actively controls bandwidth for the specific flow so that the end user device initiates switching to a different encoding/resolution, based on overall monitored network traffic. In another embodiment, the device automatically selects the appropriate format and resolution based on overall bandwidth limitations, independent of the end user.

Brief Description of the Figures

Figure 1 shows an example Cellular Operator's 3G/UMTS network, showing the 3GPP Standards defined Network Elements and the corresponding interfaces between these Network Elements;

Figure 2 is an example configuration showing the deployment of an embodiment of the invention on the IuPS interface between the RNC and SGSN in the 3GPP/UMTS Network; Figure 3 shows a block diagram of the invention in accordance with one embodiment;

Figure 4 is a diagram showing the transmission of multimedia files from a content server to user equipment according to one embodiment;

Figures 5A-B show two embodiments of a schedule file used according to the present invention.

Detailed Description of the Invention

Figure 2 shows a device 112 capable of understanding communications on a RAN network and illustrates a possible interception point where that device may be inserted in a 3G/UMTS network. The device 112 is located between the RNC 105 and the SGSN 104. This figure is an example deployment scenario in a 3G/UMTS network; while example deployments in other RAN networks, such as in a CDMA network are not shown, the methods described here are equally applicable to such networks as well. In addition, the device 112 may be located between other components in a 3G/UMTS network, such as between the Node B 106 and the RNC 105, or between the SGSN 104 and the GGSN 103.

The device 112 is capable of understanding communications between UE 107 and the core network. In some embodiments, the device 112 is able to decode the control plane and the user plane. Such a device is described in co-pending Patent Application Serial No. 12/536,537, filed August 6, 2009, the disclosure of which is incorporated herein by reference in its entirety. As such, it is able to determine when the UE 107 has requested multimedia content. Once this is known, the device 112 can optimize the delivery of that content in several ways. In one embodiment, the device 112 requests the content from the server (located on the internet) and transmits this content in a just-in-time manner to the UE 107. In another embodiment, the device 112 automatically changes or limits the options for encoding/resolution available to the client, and/or changes the delivery rate (bandwidth) for the specific multi-media flow, thereby causing the client device to initiate switching to a different resolution, based on overall monitored network and traffic to the client device. If the bandwidth available is insufficient to admit the newly requested video flow, the device 112 may block sending the video, or return error code indicating unavailability of resources, thus improving the fair allocation of network resources to other users, and not degrading the quality of experience to already started, multi-media flows. In another embodiment, the device 112 selects the appropriate format and resolution based on overall bandwidth limitations, independent of the UE 107.

Figure 3 shows a representative block diagram of the device 112. The device 112 has two interface modules 201, each of which is adapted to implement the hardware signaling required for the choice interface and the associated software protocol. This interface protocol may be IuB, IuPS or Gn, as shown in Figure 2. Each interface module 201 is adapted to receive and transmit on the selected interface. Additionally, received data is placed into a storage element 202, typically a semiconductor storage element such as a RAM, DRAM or an equivalent technology. The movement of data from the interface module to the memory 202 and vice versa may be accomplished using dedicated hardware, such as a DMA controller. Alternatively, a dedicated data movement processor may be used to handle the actual movement of data through the device 112. Once stored within the device 112, the information is processed in accordance with the RAN specifications. This may be done using dedicated control logic or a processing unit 203. The control logic/processing unit 203 may have its own local storage element 204, which contains instructions to execute and local status. This storage element may be RAM or DRAM. In addition, at least a portion of this storage element 204 may be non-volatile, such as ROM, FLASH ROM, hard disk, Solid State Disk, or the like. Using known specifications and protocols, the control logic/processing unit 203 parses the received information to understand the packet at each protocol layer. Also included may be a large storage element 205, adapted to hold cached information. In some embodiments, this cache storage may be semiconductor memory, such as RAM or DRAM. In other embodiments, this cache storage may be a rotating media, such as a disk drive or other large storage device. The control logic/processing unit may be physically implemented in a variety of technologies. For example, it may be a general-purpose processor, executing a set of instructions from an internal or external storage device .

In another embodiment, a dedicated hardware device having embedded instructions or state machines may be used to perform the functions described. Throughout this disclosure, the terms "control logic" and "processing unit" are used interchangeably to designate an entity adapted to perform the set of functions described.

The device also contains software capable of performing the functions described herein. The software may be written in any suitable programming language and the choice is not limited by this disclosure. Additionally, all applications and software described herein are computer executable instructions that are contained on a computer-readable media. For example, the software and applications may be stored in a read only memory, a rewritable memory, or within an embedded processing unit. The particular computer on which this software executes is application dependent and not limited by the present invention.

Specifically, in one embodiment, the software includes a content aware pacing algorithm. This algorithm monitors the file parameters delivered in a container file. Based on these parameters and the available RAN bandwidth, the pacing algorithm delivers multimedia data to UE 107 in such a way so as to optimize playback quality and RAN network usage.

As described above, a client or UE 107 may have a media player, such as a Flash based video player. Media players typically buffer a predetermined amount of data before the video begins playing. However, once the media player begins displaying the content, the amount of content needed by the player must be delivered to the client before it is needed for display in order for the video to be stutter free. Hence, the device 112, and specifically the content aware pacing algorithm, uses a first configuration parameter referred to as keep_ahead _r which is the amount of data, as measured in seconds, needed before starting to play the video. In other words, the video file is played in a media player, and as such, it takes a determinable time to display the content. Thus, when keep_ahead amount of data is provided, it is known that this amount of data will require keep_ahead second to display. Note that this amount of data is measured in terms of time. Therefore, depending on the video encoding and compression, two segments, each supplying keep_ahead worth of data, may comprise different numbers of bytes.

A second configuration parameter is referred to as latenc ' period, which is the Round Trip Time (RTT) from the UE 107 to the network device or server delivering the content. The round trip delay is estimated at different intervals during content delivery, or can be preconfigured. The parameter latency_period is meant to compensate for variation in network latencies to ensure stutter free delivery. Typically, the media player in the UE 107 needs keep_ahead seconds of data before starting the presentation, and needs to continue receiving that amount of data during play, so that the player does not run out of data during the play. If the network cannot keep up, i.e. cannot keep the buffer filled, the player is forced to pause and stutter during the playback of the content .

In other words, in a simple example, the device uses a keep_ahead parameter of 10 seconds and a latency_per±od of 2 seconds. In this scenario, the device 112 sends content to the UE 107 in 10 second chunks. Knowing that it may take up to 2 seconds

(i.e. latency_per±od) to transfers the next chunk to the UE 107, the device 112 may start transmitting the next chunk 8 seconds later

(i.e. keep_ahead - latency_per±od) . In order to perform this scheme, the device 112 needs to know that the requested content is indeed multimedia, and also must know the amount of data that is required for each time period.

The device and method of the current invention use the metadata stored within the container file that describes the video object and uses this information along with a predetermined amount of read- ahead buffering to ensure that the Quality of Experience (QoE) of the video is retained. In some embodiments, the container file is referred to as an .flv file. The interested reader is referred to http://www.adobe.com/devnet/flv/ for a complete description of the FLV file format. Other container file formats also exist, and the present invention is not limited to this particular format. It is important to note that the device 112 constructs the delivery schedule based on the media container-type meta-data that it monitors and decodes from the incoming stream when the video is first accessed and not yet cached.

As described above and in co-pending Patent Application Serial No. 12/536,537, the device 112 intercepts all communications between the UE 107 and the core network. It then decapsulates all traffic from the UE 107 to understand the communications that are occurring and re-encapsulates the traffic before transmitting it upstream to the core network. Similarly, the device 112 decapsulates traffic coming from the core network to understand communications back to the UE 107. As before, it re-encapsulates the data before transmitting it to the UE 107. In one embodiment, when a video download is requested by the UE 107, the device 112 may optionally parse the FLV file returned by the content server in the core network to determine the details of the video. The number of bytes to be downloaded per second is computed by the content-aware pacing algorithm located within the device 112. A brief description of the FLV file format is provided for completeness.

The FLV file contains a header and a body. The header contains information on whether there is video/audio content in the body of the file. The body of the file describes the stream and is composed of a set of tags. A tag contains the type of stream, length of data, and time of presentation of the data. For example, a tag might include that the stream is a video stream containing 99K bytes that should be presented 40 sec into the clip. Tags can be of three types: audio, video or script-data. A video tag describes a frame including the frame type, codec id, and the data. The frame type can be one of five options: Key frame, inter frame, disposable inter frame for H.263 only, a generated key frame or a command frame. Further the tag contains information about the type of codec used (JPEG, H.263, Screen video, AVC, etc.) . The content-aware pacing algorithm utilizes timestamp and data size parameters in the FLV tag. However, it can use other information such as disposable inter- frame to drop frames in case of buffer under runs. While the operation is described here uses media-container of type FLV, the current methods are applicable to other container types as well (for example MP4, 3GP, 3G2) and are not limited by this disclosure.

The content-aware algorithm parses the FLV tags for timestamp and data size parameters. For VBR (Variable Bit Rate) video, the amount of data needed to play a second of video varies. The content- aware pacing algorithm creates a transmit schedule based on the information parsed from the FLV file. It computes the amount of data that needs to be delivered per second from the timestamps for frame (FLV tag) delivery and the associated data lengths. In some embodiments, this transmit schedule is generated as the file is received from the content server, and is kept only until the content has been transmitted to the UE 107. In other embodiments, the transmit schedule, in the form of per second information, is stored in a schedule file. This schedule file may be stored in the storage element 202 of the device 112. The schedule file contains entries for the amount of video content to be served per second. The content-aware pacing algorithm uses the schedule file along with the configured buffering parameters to implement a just-in-time video content delivery to the end user.

The adaptive chunked content-aware pacing algorithm of the current invention starts by scheduling download of content data-size to meet keep_ahead seconds of play-time. Since the underlying transport is TCP, the actual delivery time for sending the scheduled chunk depends on the link-bandwidth achieved to the client. The next data chunk delivery happens at (current_time + keep_ahead — latency_per±od) . The amount of data is determined from the schedule file associated with the video object. Assuming a keep_ahead of 5 seconds, and a latency_per±od of 1 sec, the next chunk is scheduled at time=4 seconds. Subsequent chunks are scheduled 4 seconds after the transmission of the previous chunk. While the example shows burst scheduling at regular intervals, the scheduling intervals for chunks could vary depending on the current position in the media play, and relative variations of client link bandwidths when previous chunks or other multi-media, web pages, files etc. are transmitted to the same client device or other devices that share the same sector. At every burst, a minimum of keep_ahead worth of data would be sent. Each data chunk delivery happens at the maximum transfer rate that the underlying network layer supports.

As and when video content is delivered to the device 112, the video delivery schedule is built incrementally and, optionally, is written to a schedule file, examples of which are shown in Figures 5A-B. As described above, in other embodiments, the video delivery schedule is generated and maintained only until the content is transmitted to the UE 107. The schedule file contains an entry per second. In Figure 5A, each entry contains two fields: 1) play time in seconds and 2) the cumulative number of bytes of multi-media content that needs to be delivered to the end user (data _n ) . Thus, the data to be delivered in the η' ^ί1:ι second is (data _n - data _n -i ) . Once the schedule is built completely, parsing of the stream stops. This schedule can also be generated offline and kept as metadata information with the video-object or it can be built dynamically and incrementally as specified above.

Figure 4 shows a graph detailing communications between the device 112, the content server and the UE 107. As described above, the transaction begin when the UE 107 requests a multimedia file from a content server. The device 112 parses this information to understand that this is a transaction which can be paced using this content-aware pacing algorithm. The device 112 forwards the request to the content server. In response, the content server returns blocks of data, wherein each includes a header, such as a FLV format, which details specific information about the data, such as its length in bytes and its presentation time. The device 112 continues receiving this information from the content server and caches the multimedia file in its internal memory or storage element. At the same time, the device 112 is able to begin creating a schedule file, such as that shown in Figure 5A, which shows the cumulative bytes required for each time period. In other embodiments, such as that of Figure 5B, the schedule file stores the bytes required for each time interval, rather than the cumulative or running byte count.

As the device 112 receives data bytes from the multi-media content, it sends this data until keep_ahead bytes of data have been transmitted to the UE 107. Once this amount of data is transmitted to the UE 107, the device 112 pauses until the next configured chunk delivery time, as described above. The device 112 stores any bytes that it receives from the content server, such as in cache 205, for future transmissions to the UE 107. The device 112 continues to receive data from the content server and continues building the schedule file. At the specified time, such as keep_ahead latency_per±od seconds after transmission of the previous transfer, the device 112 sends a second packet of keep_ahead bytes of data to the UE 107. This continues until either the file is completely transferred, or the user navigates away from the video. This latter action is detected by the device 112 typically by the termination of the TCP connection to the UE 107. Note that the device 112 interfaces on one side (i.e. to/from the content server) which is typically wireline and therefore has more predictable bandwidth and latency characteristics. The device 112 also interfaces to the UE 107, where wireless communication and varying numbers of devices and usage patterns impact latency. Thus, the device 112 serves to separate these two interfaces, allowing the content server to deliver data at one rate, and pacing the data to the UE 107 at a second rate.

While Figure 4 shows that the device 112 receives content from the content server, it should be noted that the pacing algorithm can be used with previously cached content. In this embodiment, the device 112 has a copy of the desired content, and has a previously constructed schedule file. Using this information, the pacing algorithm is able to transmit the content to the UE 107 as described above .

The determination of the keep_ahead and latency_per±od parameters may be done in a variety of ways. In some embodiments, these values are fixed and remain constant for all UEs . In other embodiments, the parameters vary based on network conditions. For example, the latency_per±od is most obviously related to overall network bandwidth. The device 112 can monitor actual RAN bandwidth and adjust the latenc _period parameter in real time. In other embodiments, this parameter may vary based on other characteristics, such as time of day. For example, the device 112 may decrease the latency_per±od during the late night hours, since the expected network traffic is lower. In other embodiments, the latency_period may be based on previously monitored bandwidth to this UE 107.

Similarly, the keep_ahead parameter can be modified in a number of ways. The choice of an optimal value for the keep_ahead parameter is based on balancing two competing goals. On one hand, smaller keep_ahead values allow the video to begin playing on the UE 107 sooner. These small values also minimize wasted bandwidth in the event that the user navigates away from the video. On the other hand, larger keep_ahead values are less susceptible to abrupt changes in network bandwidth, thereby minimizing the probability of pausing or stuttering. As suggested above, monitored network bandwidth, time of day, previously observed bandwidth or other characteristics can be used to set or modify the keep_ahead parameter .

There could be possibility of buffer under-run wherein server is not able to send out keep_ahead worth of data in specified time interval, in which case content-aware pacing algorithm sends out data at whatever rate that the underlying transport layer supports, and pacing benefits will not be seen here. However optionally information about disposable inter frame can be used here to drop inter-frames in case of under run.

In another embodiment, the device 112 is able to modify and select video formats and resolutions based on network activity. Again, since the device 112 is able to intercept and decapsulate user and control planes, it is able to determine the content of requests going from the UE 107 to a content server. Thus, the device 112 is able to modify these responses or the replies to these requests as appropriate. Using this technique, the device 112 is able to achieve network controlled bit rate selection. This allows the device 112 to set the initial resolution, which is referred to as Network Controlled Bit Rate selection.

In one embodiment, the UE 107 requests a video clip from a content server, such as YouTube . In response, the content server may provide a listing of the different resolutions in which that video is available. This list may be referred to as a format MAP. Typically, this list is supplied to the UE 107, which then selects a particular resolution. However, the device 112, having knowledge of this transaction may modify the format map returned to the US 107, based on observed network bandwidth. For example, the client may request a video which is available in a plurality of formats, from 240 to 720p. In response to a request from the client, the content provider will supply this full list of formats. The device 112 may, based on observed bandwidth, determine that it is not possible to display the 720p version of this video without pauses or stutters. Therefore, before sending this response back to the client, the device 112 modifies the response, and removes any formats which require bandwidth in excess of that which is currently available. Thus, the response received by the client contains only those formats which the device 112 believes can be displayed with an acceptable QoE .

In another embodiment, the request from the UE 107 may contain a flag, such as "&hd=l", indicating that the content server should include high definitions versions of the video in the format map. The device 112, determining that a high definition video cannot be properly displayed, may modify the request before transmitting it to the content server, such as by removing the flag "&hd=l". In this way, the content server will only return those formats which are not high definition.

Therefore, the device 112 is able to modify the format map received by the UE 107, either by modifying the request to the content server, or by modifying the response from the content server .

In another embodiment, the device 112 modifies the resolution of the video during transmission. For example, smooth streaming or adaptive streaming over HTTP is known. In this mechanism, the client continues to monitor its perceived network bandwidth, and adjusts the resolution or format of the video being played in real time. In other words, the client may be playing a 720p version of a video and detect that the bandwidth of the network has dramatically dropped. In response, it may, without closing the media player, request a lower resolution version of the video, thereby changing resolution in mid-stream. Variations of this technique exist in Microsoft Silverlight, Netflix player and Adobe OSMF . However, in all of these embodiments, it is the client that determines the available bandwidth. Therefore, the client is typically reactive to changes (both positive and negative) in bandwidth.

RAN networks offer unique challenges in that the number of devices in the network and the amount of traffic each generates can change rapidly and unpredictably. Thus, it would be beneficial if a component, such as device 112, were able to monitor the network activity and make decisions about changing video resolutions. In one embodiment, the device 112 transmits video data to the UE 107. This can be done according to the technique described above, or using other traditional methods. While this transmission of the video is occurring, the device 112 continuously monitors network activity. If a change is noted in network bandwidth (either positive or negative), the device 112 can adapt.

In one embodiment, the device 112, upon determining a change in RAN bandwidth, begins requesting the alternative resolution from the content server (if it is not cached) . Once the device 112 has received this video in the alternative resolution, it automatically begins transmitting it to the UE 107. The device 112 may recognize a convenient boundary point (for example, key frames or at the GOP Frame boundaries) for this transition. This embodiment assumes that the media player of the UE 107 is able to adapt to changes in resolution which are made unilaterally by the device 112.

In other embodiments, the media player of the UE 107 may only be able to adapt to changes in resolution if those changes were requested by the media player itself. In this embodiment, the device 112 may modify its behavior so as to encourage the media player to change resolutions. For example, assume that the RAN bandwidth greatly increased. In order to get the media player to request a higher resolution format of the video, the device 112 may accelerate its delivery of video data to the UE 107, so as to fill or overcome its buffer. This increased data rate to the UE 107 may cause the media player to request a higher resolution format. Conversely, the device 112 may slow its delivery of data to the UE 107 so that the media player's buffer empties or nearly empties, causing the media player to request a lower resolution version.

If these techniques are used in conjunction with the initial resolution selection, the network is able to determine and optimize video resolutions throughout the RAN, based on actual monitored traffic and congestion.

In another embodiment, the device 112 parses the communications between the UE 107 and the core network to determine that the UE 107 is requesting a multi-media file. The device 112, aware of the bandwidth required by such a multi-media file, first determines the available bandwidth of the RAN. For example, if there is little other traffic, the device 112 will honor the request. However, if the available bandwidth is low, due to a high number of devices, a large number of data-intensive transactions, or a combination of these factors, the device 112 may deny the request of the UE 107. In this case, the device 112 may return an error message or simply not deliver the file to the UE 107 if the available bandwidth is above a predetermined threshold. In some embodiments, the device 112 denies the request before the request for data is forwarded to the content server. In other embodiments, the device 112 makes the determination to deny the request after receiving at least part of the file from the content server, but prior to delivering the content to the UE 107.

The terms and expressions which have been employed herein are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described (or portions thereof) . It is also recognized that various modifications are possible within the scope of the claims. Other modifications, variations, and alternatives are also possible. Accordingly, the foregoing description is by way of example only and is not intended as limiting.