Anti-alignment download orchestration method based on multi-resource hybrid burst and multiplexing

Abstract

This application discloses an anti-alignment download orchestration method based on multi-resource hybrid bursts and multiplexing, comprising: intercepting a set of resource requests to be downloaded initiated by a client, the set of resource requests to be downloaded including at least one target media stream resource request and at least one auxiliary resource request; calculating the latest start time corresponding to each resource request according to the playback dependency relationship and estimated bandwidth of each resource, so as to meet the condition that the playback buffer is not lower than the safety level; dividing the video segment data corresponding to the target media stream resource request into multiple video data blocks, and generating a frame-level interleaved transmission sequence according to the latest start time, wherein the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource requests in the frame-level interleaved transmission sequence are alternately arranged or aggregated on the time axis to form a hybrid burst plan; and sending the corresponding request frame and data frame to the server according to the frame-level interleaved transmission sequence through different stream identifiers of the transmission protocol supporting multiplexing.

Description

Technical Field

[0001] This application relates to the field of computer network transmission technology, and more specifically, to an anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing. Background Technology

[0002] Next-generation internet transport protocols have been widely applied in streaming media transmission scenarios. These protocols introduce the core feature of multiplexing, allowing multiple logical streams to be transmitted simultaneously over a single physical connection, thereby effectively reducing connection establishment overhead and head-of-line congestion. In modern streaming media applications, when a client plays video, it not only needs to load video stream data but also concurrently load various types of resources such as audio streams, subtitle files, thumbnails, recommendation lists, and application programming interface responses. These resources share the same physical connection at the transport layer, theoretically enabling efficient parallel transmission.

[0003] However, existing browsers and network libraries typically employ request / response models or priority-based serialization models in their scheduling strategies, which have the following technical drawbacks: (1) Logical isolation leads to leakage of traffic characteristics: Although different types of resources share the same connection at the physical layer, due to the head-of-line blocking control of the transmission control protocol or the limitations of the application layer scheduling logic, video streams (usually given high priority) often preempt most of the bandwidth, forming a huge, continuous data burst. Meanwhile, low-priority resources such as audio streams and images are squeezed into the gaps in video transmission, or can only start transmitting after the video transmission is completed. This scheduling method results in a clear distinction in the traffic of different types of resources on the timeline.

[0004] (2) Attackers can easily separate mixed traffic: For network traffic analysis attackers, the above scheduling method exposes a clear traffic fingerprint. Attackers can easily separate the originally mixed traffic by using simple threshold filtering methods, such as classifying bursts greater than 500KB as video streams and bursts less than 100KB as audio streams. Then, they can perform fingerprint identification on the separated video streams, which seriously threatens the privacy and security of users.

[0005] (3) Existing schedulers lack proactive obfuscation mechanisms: Existing schedulers only focus on quality of service metrics during design, such as downloading video data as quickly as possible to avoid playback stuttering, and have never considered using concurrent non-video traffic as a natural cover to obfuscate video fingerprints. Although some studies have attempted to mask real traffic characteristics by injecting junk traffic, this method increases additional bandwidth consumption and power consumption, which is not conducive to practical deployment.

[0006] In summary, designing a scheduling method that proactively counters traffic alignment analysis while ensuring the player buffer does not run out (i.e., does not cause stuttering) under the premise of fully utilizing the concurrent transmission characteristics of multiplexing protocols has become a pressing technical problem in this field. Summary of the Invention

[0007] This invention provides an anti-alignment download orchestration method based on multi-resource hybrid bursts and multiplexing, which solves the technical problem in the prior art where the traffic characteristics of video streams and auxiliary resources are easily separated and identified by attackers in a multiplexed transmission environment, leading to the leakage of user privacy. Without violating the player buffer constraints (i.e. without causing stuttering), the video stream data frames are deeply interwoven with the data frames of other resources such as audio and images on the time axis, making it impossible for attackers to extract pure video features from the mixed traffic.

[0008] To achieve the above objectives, according to a first aspect of the present invention, an anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing is provided. The method includes: intercepting a set of resource requests to be downloaded initiated by a client, the set of resource requests including at least one target media stream resource request and at least one auxiliary resource request, the target media stream resource request being a video stream segmentation request, and the auxiliary resource requests including audio stream requests, image requests, subtitle requests, or interface response requests; calculating the latest start time corresponding to each resource request based on the playback dependencies and estimated bandwidth to ensure that the playback buffer is not lower than a safety level; dividing the video segmentation data corresponding to the target media stream resource request into multiple video data blocks, and generating a frame-level interleaved transmission sequence based on the latest start time, wherein the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource requests in the frame-level interleaved transmission sequence are alternately arranged or aggregated on the time axis to form a hybrid burst plan; and sending corresponding request frames and data frames to the server according to the frame-level interleaved transmission sequence using different stream identifiers of a transmission protocol supporting multiplexing.

[0009] In an exemplary embodiment, generating a frame-level interleaved transmission sequence includes: within a planned time window, searching for an interleaved sequence in which video data blocks and auxiliary data blocks are alternately arranged in transmission order, with the optimization objective of maximizing the mixing entropy of the output traffic sequence or minimizing the self-similarity of the output traffic sequence, and at least one auxiliary data block is inserted between any two consecutive video data blocks; wherein, the planned time window is determined based on the latest start time and the current time, the mixing entropy is used to characterize the degree of mixing of data frames of different resource types in the output traffic sequence, and the auxiliary data block is obtained by dividing the resource data corresponding to the auxiliary resource request according to a set granularity.

[0010] In an exemplary embodiment, the method further includes: calculating the mixing entropy of the current output traffic sequence in real time; when the mixing entropy is lower than a preset mixing degree threshold, adjusting the frame-level interleaved transmission sequence, including at least one of the following methods: delaying the transmission of some video data blocks, sending auxiliary data blocks in advance, changing the interleaving density of data blocks, or inserting camouflage data blocks.

[0011] In an exemplary embodiment, the auxiliary data block is obtained by dividing the resource data corresponding to the auxiliary resource request according to a set granularity, specifically including: obtaining the set granularity of the video data block as the target granularity; dividing the resource data corresponding to the auxiliary resource request into multiple auxiliary data blocks, such that the size of each auxiliary data block matches the target granularity; wherein, the auxiliary resources include image resources, and the image resources are divided into multiple image data blocks corresponding to the number of video data blocks.

[0012] In one exemplary embodiment, the method further includes: real-time monitoring of the current data volume in the player buffer; when the current data volume is lower than a preset safety threshold, exiting the execution of the hybrid burst plan; prioritizing the scheduling of sending video segment data corresponding to the target media stream resource request until the current data volume recovers to above the safety threshold, and then resuming the execution of the hybrid burst plan.

[0013] In an exemplary embodiment, sending corresponding request frames and data frames through different stream identifiers includes: opening multiple logical streams and assigning different stream identifiers; sequentially encapsulating each data block into a data frame with the corresponding stream identifier and writing it into the underlying transmission buffer according to the frame-level interleaved transmission sequence; when two adjacent data frames belong to different resource types and have different data frame lengths, inserting a padding frame between the two data frames to align the data frames of different resources on the transmission time axis.

[0014] According to a second aspect of the present invention, an anti-alignment download orchestration device based on multi-resource hybrid bursting and multiplexing is also provided. The device includes: a resource set management module, used to intercept a set of resource requests to be downloaded initiated by a client, the set of resource requests to be downloaded including at least one target media stream resource request and at least one auxiliary resource request, the target media stream resource request being a video stream segmentation request, and the auxiliary resource requests including audio stream requests, image requests, subtitle requests, or interface response requests; and a dependency constraint module, used to calculate the corresponding resource requests based on the playback dependencies of each resource and the estimated bandwidth. The latest start time is set to ensure that the playback buffer does not fall below the safety level. The hybrid orchestration module is used to divide the video segment data corresponding to the target media stream resource request into multiple video data blocks, and generate a frame-level interleaved transmission sequence according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are arranged alternately or aggregated on the time axis to form a hybrid burst plan. The concurrent scheduling module is used to send the corresponding request frames and data frames to the server according to the frame-level interleaved transmission sequence and through different stream identifiers of the transmission protocol that supports multiplexing.

[0015] In one exemplary embodiment, the apparatus further includes: an evaluation feedback module, configured to calculate the mixing entropy of the current output traffic sequence in real time; and when the mixing entropy is lower than a preset mixing degree threshold, to adjust the frame-level interleaved transmission sequence to increase the mixing entropy of the current output traffic sequence.

[0016] According to a third aspect of the invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is configured to execute the above-described anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing at runtime.

[0017] According to a fourth aspect of the present invention, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-described anti-alignment download orchestration method based on multi-resource hybrid burst and multiplexing via the computer program.

[0018] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: (1) This invention provides an anti-alignment download orchestration method based on multi-resource hybrid bursts and multiplexing. It deeply interweaves video stream data frames with data frames from auxiliary resources such as audio and images on the time axis, making it impossible for attackers to observe the mixed traffic statistics and to easily distinguish the video stream from the background stream based on the size or time interval of the data bursts. Attackers cannot extract pure video features from the mixed traffic, thus resisting traffic alignment analysis attacks and protecting user viewing privacy. Furthermore, it monitors the data volume in the player buffer in real time, prioritizing the download of video data when the buffer is tight, ensuring that hybrid burst scheduling is performed without causing playback stuttering. This method balances privacy protection and user experience, avoiding the problem of sacrificing playback smoothness for excessive obfuscation effects.

[0019] (2) This invention utilizes legitimate resources that would otherwise need to be downloaded (such as audio, images, subtitles, etc.) as a cover, achieving privacy protection without increasing total bandwidth consumption, avoiding additional bandwidth usage and power consumption, making this solution more economical and feasible in actual deployment. The client only needs to deploy the scheduling device of this invention at the application layer or transport layer to achieve anti-alignment download orchestration, which has low deployment cost, strong practicality, and can be applied to existing streaming media service platforms. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating an optional anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing provided for embodiments of this application; Figure 2 A flowchart illustrating another optional anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing provided for embodiments of this application; Figure 3 A schematic diagram of an optional anti-alignment download orchestration device based on multi-resource hybrid burst and multiplexing provided for embodiments of this application; Figure 4 This is a schematic diagram of an optional electronic device provided in an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0023] The terms "first," "second," "third," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0024] According to one aspect of the embodiments of this application, an anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing is provided. The following is in conjunction with... Figure 1 This application describes an anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing provided in its embodiments.

[0025] Figure 1 This is a flowchart illustrating an optional anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing provided in an embodiment of this application, as shown below. Figure 1 As shown, the process includes the following steps: S102, intercept the set of resource requests to be downloaded initiated by the client. The set of resource requests to be downloaded includes at least one target media stream resource request and at least one auxiliary resource request. The target media stream resource request is a video stream segmentation request. The auxiliary resource request includes an audio stream request, an image request, a subtitle request, or an interface response request. S104. Based on the playback dependencies of each resource and the estimated bandwidth, calculate the latest start time corresponding to each resource request to ensure that the playback buffer is not lower than the safety level. S106, the video segment data corresponding to the target media stream resource request is divided into multiple video data blocks, and a frame-level interleaved transmission sequence is generated according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are alternately arranged or aggregated on the time axis to form a hybrid burst plan. S108, according to the frame-level interleaved transmission sequence, send the corresponding request frame and data frame to the server through different stream identifiers of the transmission protocol that supports multiplexing.

[0026] This application provides an anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing, which can be applied to scenarios where video streams and non-video resources are downloaded concurrently on the same multiplexed connection and there is a risk of traffic privacy leakage.

[0027] It should be noted that this embodiment provides an anti-alignment download orchestration method based on multi-resource hybrid burst and multiplexing. This method can be applied to the client side, where the client establishes a connection with the server through a multiplexing-supporting transmission protocol and downloads various types of network resources.

[0028] Combination Figure 1 and Figure 2 As shown, the client intercepts its own set of resource requests to be downloaded. This set includes at least one target media stream resource request and at least one auxiliary resource request. Here, the target media stream resource request is a video stream segment request, such as video segment data requested in a streaming media player. Auxiliary resource requests include audio stream requests, image requests, subtitle requests, or API response requests, such as audio segments corresponding to video segments, video thumbnails, subtitle files, and recommendation list application programming interface (API) responses.

[0029] By intercepting and classifying resource requests initiated by clients, it is possible to identify the target media stream resources (video) that need to be protected, as well as the shield resources (auxiliary resources) used to obfuscate traffic characteristics.

[0030] Optionally, the latest start time for each resource request is calculated based on the playback dependencies of each resource and the estimated bandwidth. Specifically, according to the formula:

[0031] Calculate the latest start time for each resource request. ,in, This is the latest deadline for completing this resource request. The data size of this resource. This represents the estimated network bandwidth.

[0032] Ensure that the playback buffer level remains above the safe threshold, and rationally schedule the download sequence of each resource. For segmented video resources, the deadline is determined by the player's buffer state and must be completed before the buffer is exhausted to avoid playback stuttering. For auxiliary resources, the deadline may be more lenient or have no strict deadline requirements.

[0033] Furthermore, the video segment data corresponding to the target media stream resource request is divided into multiple video data blocks. For example, a 1MB video segment can be divided into 10 100KB video data blocks.

[0034] Then, based on the calculated latest start time, a frame-level interleaved transmission sequence is generated. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource requests are arranged alternately or aggregated on the time axis to form a hybrid burst plan.

[0035] In a preferred embodiment, the specific method for generating the frame-level interleaved transmission sequence is as follows: Within a planned time window, with the optimization objective of maximizing the mixing entropy of the output traffic sequence or minimizing the self-similarity of the output traffic sequence, an interleaved sequence is searched to obtain one in which video data blocks and auxiliary data blocks are alternately arranged in transmission order, and at least one auxiliary data block is inserted between any two consecutive video data blocks. The planned time window is determined based on the latest start time and the current time; the mixing entropy is used to characterize the degree of mixing of data frames of different resource types in the output traffic sequence; and the auxiliary data blocks are obtained by dividing the resource data corresponding to the auxiliary resource requests according to a set granularity.

[0036] For example, suppose there are video segments in the waiting queue. (Size 1MB, latest deadline for this resource request is 2s), audio segments (Size 100KB, latest deadline for this resource request is 2 seconds), Image (Size 200KB, no latest completion deadline). Understandably, traditional scheduling would send it first. Send all data again Finally send The line traffic characteristics are [1MB Burst] ... [100KB Burst] ... [200KB Burst]. Attackers can easily identify the first large burst as video.

[0037] Using the hybrid scheduling method of this application, Split into 10 100KB data blocks: ,Will Split into two 100KB data blocks: Generate an interleaved sequence: - - - - The line characteristics form a continuous, long-tailed, mixed burst containing data frames from different streams. Attackers cannot accurately infer the size of video segments based on the burst size, nor can they infer the boundaries of video segments based on the burst intervals.

[0038] For example, according to the generated frame-level interleaved transmission sequence, corresponding request frames and data frames are sent to the server using different stream identifiers of multiplexed transport protocols. Multiplexed transport protocols include Hypertext Transfer Protocol Version 2 (HTTP / 2) or Hypertext Transfer Protocol version 3 over Quick UDP Internet Connections (HTTP / 3 over QUIC).

[0039] Specifically, firstly, multiple logical streams are opened and different stream identifiers (Stream IDs) are assigned. Then, according to the frame-level interleaved transmission sequence, each data block is encapsulated into a data frame with the corresponding stream identifier and written to the underlying transmission buffer. When two adjacent data frames belong to different resource types and have different lengths, a padding frame is inserted between the two data frames to align the data frames of different resources on the transmission timeline.

[0040] In this embodiment, video stream data frames are deeply interwoven with data frames from auxiliary resources such as audio and images on the time axis. This makes it impossible for attackers to observe the mixed traffic statistics, and they cannot easily distinguish between the video stream and the background stream by the size or time interval of the data bursts. This effectively resists traffic alignment analysis attacks and protects the user's privacy and security.

[0041] Through steps S102 to S108, the process involves intercepting a set of resource requests initiated by the client, including at least one target media stream resource request and at least one auxiliary resource request. Based on the playback dependencies and estimated bandwidth of each resource, the latest start time for each resource request is calculated to ensure the playback buffer remains above a safe level. The video segment data corresponding to the target media stream resource request is divided into multiple video data blocks, and a frame-level interleaved transmission sequence is generated based on the latest start time. In this sequence, video data blocks and auxiliary data blocks corresponding to auxiliary resource requests are alternately or aggregated on the timeline, forming a hybrid burst plan. Following the frame-level interleaved transmission sequence, corresponding request frames and data frames are sent to the server using different stream identifiers from multiplexed transmission protocols. This approach resists traffic alignment analysis attacks, balances privacy protection and user experience, avoids sacrificing playback smoothness for excessive obfuscation effects, and boasts low deployment costs and high practicality.

[0042] In one exemplary embodiment, generating the frame-level interleaved transmission sequence includes: Within the planning time window, with the optimization objective of maximizing the mixed entropy of the output traffic sequence or minimizing the self-similarity of the output traffic sequence, an interleaved sequence is searched to ensure that video data blocks and auxiliary data blocks are alternately arranged in the transmission order, and that at least one auxiliary data block is inserted between any two consecutive video data blocks. The planning time window is determined based on the latest start time and the current time. The mixed entropy is used to characterize the degree of mixing of data frames of different resource types in the output traffic sequence. The auxiliary data block is obtained by dividing the resource data corresponding to the auxiliary resource request according to a set granularity.

[0043] Optionally, when generating this hybrid contingency plan, a planning time window is first determined. , ].in, The current moment; To obtain the maximum value of the latest start time among all pending resource requests, the length of the planning time window is determined. This window defines the time boundaries for this hybrid orchestration: all video data blocks and auxiliary data blocks must be scheduled for transmission within this window, while also satisfying their respective latest start time constraints.

[0044] The target media stream (video segment) is divided into multiple video data blocks of a fixed size (e.g., 100KB). Auxiliary resources (audio, images, subtitles, etc.) are divided according to their own data size and a set granularity. If the size of the auxiliary resource data is divisible by the set granularity, it is evenly split; if it is not divisible, the last block is allowed to be smaller than the set granularity, or it can be merged with subsequent resources to make up the size. In this embodiment, the set granularity is preferably equal to the fixed size of the video data blocks (e.g., all set to 100KB) to maximize the alignment of the staggered arrangement.

[0045] Within a planning time window, a discrete sorting problem is constructed. The decision variable is the transmission order sequence, where each element is either a video data block or an auxiliary data block. Constraints are: the actual transmission time of each data block must not be later than the latest start time of its corresponding resource; at least one auxiliary data block must be inserted between any two consecutive video data blocks; video data blocks and auxiliary data blocks are alternately arranged or aggregated on the time axis (aggregation means allowing a small number of consecutive blocks of the same type, but the overall mixing degree meets the lower limit). The optimization objective is to maximize the mixing entropy or minimize self-similarity. The client solves this optimization problem within the planning window to obtain the optimal (or near-optimal) staggered sequence.

[0046] In an exemplary embodiment, the auxiliary data block is obtained by dividing the resource data corresponding to the auxiliary resource request according to a set granularity, specifically including: The set granularity of the video data block is obtained as the target granularity; The resource data corresponding to the auxiliary resource request is divided into multiple auxiliary data blocks, such that the size of each auxiliary data block matches the target granularity. The auxiliary resources include image resources, and the image resources are divided into multiple image data blocks corresponding to the number of video data blocks.

[0047] In this embodiment, the following preferred strategy can be adopted for the division of image resources: Uniform splitting: The image resource is evenly split into multiple image data blocks, with the size of each image data block matching the target granularity of the video data block. For example, if the image size is 400KB and the target granularity is 100KB, it will be split into four 100KB image data blocks.

[0048] Quantity-matching splitting: In a preferred embodiment, image resources are divided into multiple image data blocks corresponding to the number of video data blocks. For example, if a video segment is divided into 10 video data blocks, then the images are also divided into 10 image data blocks (even if the size of each image data block may be smaller than the target granularity, it is padded with frames). This splitting method can achieve a one-to-one correspondence and interleaved arrangement of video data blocks and image data blocks, maximizing the hybrid entropy.

[0049] Adaptive partitioning: Based on the total amount of available image resources and the number of video data blocks, the granularity and number of image data blocks are adaptively determined. For example, if multiple image resources are available, they can be merged and then partitioned uniformly to ensure a sufficient number of auxiliary data blocks for interleaving.

[0050] In one exemplary embodiment, the method further includes: Calculate the mixed entropy of the current output flow sequence in real time; When the mixing entropy is lower than a preset mixing degree threshold, the frame-level interleaved transmission sequence is adjusted, including at least one of the following methods: delaying the transmission of some video data blocks, sending auxiliary data blocks in advance, changing the interleaving density of data blocks, or inserting camouflage data blocks.

[0051] In this embodiment, while the client sends data frames to the server according to the frame-level interleaved transmission sequence, it samples the completed output traffic sequence at fixed time intervals (e.g., every 100 milliseconds or after every 10 data frames are sent) and calculates its mixed entropy in real time.

[0052] For example, if video frames account for 50%, audio frames account for 25%, and image frames account for 25% within the sampling window, then the mixing entropy H = (0.5log0.5+0.25log0.25+0.25log0.25)≈1.04. If video frames account for 90%, audio frames account for 5%, and image frames account for 5%, then the mixing entropy H≈0.52, indicating a low degree of mixing.

[0053] The client pre-sets a mixing threshold (e.g., 0.85), which can be dynamically adjusted based on factors such as network environment and user privacy protection requirements. When the real-time calculated mixing entropy is lower than the preset mixing threshold, it indicates that the characteristics of the video stream in the current output traffic sequence are too prominent. Attackers may be able to identify the existence and boundaries of the video stream through simple statistical analysis methods, at which point an adjustment mechanism needs to be triggered.

[0054] When the mixed entropy is detected to be below a threshold, the client immediately adjusts the frame-level interleaved transmission sequence of the remaining untransmitted portion. The adjustment method includes at least one of the following: (1) Delay the transmission of some video data blocks, postponing the transmission of some untransmitted video data blocks from the current transmission position to make room for inserting more auxiliary data blocks. It should be noted that the delayed transmission cannot violate the latest start time constraint corresponding to the video data block. During the adjustment, it will work with the dependency constraint module to ensure that the adjusted sequence still meets the condition that the playback buffer is not lower than the safety level.

[0055] (2) Send auxiliary data blocks in advance. Send auxiliary data blocks (such as images, subtitles, API responses, etc. that have not yet been requested) that were originally planned to be sent in a later time window to the current position in order to fill the gaps between video data blocks.

[0056] (3) Change the interleaving density of the data blocks, adjust the ratio of video data blocks to auxiliary data blocks, and increase the insertion frequency of auxiliary data blocks. For example, change the original insertion of 1 auxiliary block for every 3 video blocks sent to insertion of 1 auxiliary block for every 1 video block sent, thereby increasing the mixing entropy.

[0057] (4) Inserting camouflage data blocks: When the total amount of existing auxiliary resource data is insufficient to support the required level of mixing, the client can request pre-configured camouflage resources from the server, such as requesting a harmless transparent image or empty data interface that will not be rendered or displayed; or directly read previously downloaded camouflage data blocks from the local cache and insert them into the current transmission sequence. These camouflage data blocks are not used at the application layer and only play a role in obfuscating traffic characteristics at the transport layer.

[0058] After the adjustment is completed, the client continues to send data frames according to the updated frame-level interleaved transmission sequence and continuously monitors the change in mixed entropy. If the mixed entropy recovers to above the threshold, the current sequence continues to be executed; if the mixed entropy is still below the threshold, the adjustment is triggered again, forming a continuous closed-loop feedback mechanism.

[0059] Through the aforementioned dynamic feedback mechanism, the client can perceive the obfuscation effect of the current output traffic in real time during runtime, and proactively adjust the sending strategy when the obfuscation effect is poor, ensuring that the traffic observed by the network side maintains a high mixing entropy throughout the entire video playback process, thereby continuously and effectively resisting traffic alignment analysis attacks.

[0060] In one exemplary embodiment, the method further includes: Real-time monitoring of the current data volume in the player's buffer; When the current data volume is lower than a preset security threshold, the execution of the hybrid burst plan is terminated. Prioritize scheduling and sending the video segment data corresponding to the target media stream resource request until the current data volume recovers to above the safety threshold, then resume execution of the hybrid burst plan.

[0061] In this embodiment, while sending data frames to the server according to a frame-level interleaved transmission sequence, the current data volume of the player buffer is obtained in real time at fixed time intervals (e.g., every 200 milliseconds or after each new video data is received). The player buffer refers to the local storage area on the client used to temporarily store downloaded but not yet decoded and played video data.

[0062] The client pre-sets a safety threshold (e.g., 2 seconds of buffered data), which can be dynamically adjusted based on factors such as network conditions, video bitrate, and user-defined playback modes. When the real-time monitored buffer data volume falls below the preset safety threshold, it indicates that the player is at risk of stuttering. In this case, the hybrid burst plan needs to be immediately paused to prioritize the download of video data. The client immediately exits the current hybrid burst plan, specifically by: pausing the data frame transmission scheduling according to the frame-level interleaved transmission sequence; canceling or delaying currently waiting auxiliary resource requests (such as audio, images, subtitles, API responses, etc.); and switching the scheduling mode to pure video priority mode.

[0063] After exiting the hybrid burst plan, the client enters emergency backup mode, in which: all available bandwidth resources are prioritized for downloading the target media stream (video segments); video data blocks are no longer split and interleaved, and are sent continuously in units of the original video segments; all non-essential auxiliary resource requests are suspended, and only video segment requests are retained.

[0064] In emergency backup mode, the client continuously monitors changes in the buffer data volume. When the buffer data volume recovers to above the safety threshold (e.g., the buffer data volume reaches more than 3 seconds), it indicates that the player is out of risk of stuttering. At this time, the client: exits emergency backup mode; restarts the hybrid orchestration module, regenerates the frame-level interleaved transmission sequence based on the latest resource request set and the latest start time; resumes execution of the hybrid burst plan, and continues to send data frames of various resources in a hybrid orchestration manner.

[0065] This embodiment achieves a balance between prioritizing privacy protection (anti-alignment) and consistently maintaining smooth playback as the highest priority. When the player faces the risk of stuttering, obfuscation effects are immediately sacrificed to ensure user experience; once the buffer returns to a safe level, the obfuscation strategy is reactivated. This mechanism ensures that in practical deployment, the invention will not negatively impact the user's normal viewing experience due to an excessive pursuit of privacy protection.

[0066] In one exemplary embodiment, sending the corresponding request frame and data frame through different stream identifiers includes: Start multiple logical streams and assign different stream identifiers; According to the frame-level interleaved transmission sequence, each data block is encapsulated into a data frame with the corresponding stream identifier and written into the underlying transmission buffer. When two adjacent data frames belong to different resource types and have different lengths, a padding frame is inserted between the two data frames to align the data frames from different resources on the transmission timeline.

[0067] In this embodiment, after establishing a multiplexed transport protocol connection (such as an HTTP / 2 connection or an HTTP / 3 over QUIC connection) with the server, the client initiates an independent logical stream for each resource or data block based on the different resource types involved in the frame-level interleaved transmission sequence, and assigns a unique stream identifier (Stream ID). Each logical stream is initiated through an independent header frame, carrying the corresponding resource request Uniform Resource Locator (URL) and header information.

[0068] The client sequentially traverses each data block in the generated frame-level interleaved transmission sequence, encapsulates each data block into a data frame with the corresponding stream identifier, and writes it to the underlying transmission buffer. During the writing process, the client strictly follows the order of the frame-level interleaved transmission sequence to ensure that data frames with different stream identifiers are arranged in a predetermined order in the transmission buffer, thus presenting a mixed traffic characteristic at the physical channel output.

[0069] During the writing of data frames to the underlying transmit buffer, the client checks the attributes of two adjacent data frames. One or more padding frames are inserted between the two adjacent data frames when the following two conditions are met: Condition 1: The resource types of two adjacent data frames are different. For example, the first data frame belongs to a video stream (Stream ID=1), and the second data frame belongs to an audio stream (Stream ID=3).

[0070] Condition 2: The lengths of two adjacent data frames are different. For example, the length of the first video data frame is 100KB, and the length of the second audio data frame is 50KB.

[0071] When both of the above conditions are met, the client inserts a padding frame before the next data frame. The length of the padding frame is the difference between the lengths of the two data frames (i.e., 50KB), so that the two data frames appear as consecutive data blocks of the same length on the transmission timeline.

[0072] Through the aforementioned padding frame insertion mechanism, the length differences of data frames of different resource types on the transmission time axis are eliminated, and the data frame sequence observed by the network side appears as a uniform and continuous data stream, thereby further enhancing the effect of traffic obfuscation.

[0073] According to another aspect of the embodiments of this application, an anti-alignment download orchestration apparatus based on multi-resource hybrid burst and multiplexing is also provided for implementing the above-described anti-alignment download orchestration method based on multi-resource hybrid burst and multiplexing, comprising: The resource collection management module is used to intercept the set of resource requests to be downloaded initiated by the client. The set of resource requests to be downloaded includes at least one target media stream resource request and at least one auxiliary resource request. The target media stream resource request is a video stream segmentation request, and the auxiliary resource request includes an audio stream request, an image request, a subtitle request, or an interface response request. The dependency constraint module is used to calculate the latest start time corresponding to each resource request based on the playback dependency relationship of each resource and the estimated bandwidth, so as to meet the condition that the playback buffer is not lower than the safety level. The hybrid orchestration module is used to divide the video segment data corresponding to the target media stream resource request into multiple video data blocks, and generate a frame-level interleaved transmission sequence according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are alternately arranged or aggregated on the time axis to form a hybrid burst plan. The concurrent scheduling module is used to send corresponding request frames and data frames to the server according to the frame-level interleaved transmission sequence and through different stream identifiers of the transmission protocol that supports multiplexing.

[0074] In one exemplary embodiment, the apparatus further includes: The evaluation feedback module is used to calculate the mixed entropy of the current output flow sequence in real time; When the mixing entropy is lower than a preset mixing degree threshold, the frame-level interleaved transmission sequence is adjusted to increase the mixing entropy of the current output traffic sequence.

[0075] like Figure 3 As shown, the resource collection management module is deployed in the client's network stack, located between the application layer and the transport layer. When a client application (such as a streaming media player) initiates a Hypertext Transfer Protocol (HTTP) request, the resource collection management module first intercepts these requests and classifies them: target media stream resource requests (identified as video stream segmentation requests, marked as core resources requiring protection); and auxiliary resource requests (identified as audio stream requests, image requests, subtitle requests, or interface response requests, marked as cover resources that can be used for obfuscation). The resource collection management module maintains a queue of resource requests to be sent, recording the type, priority, and deadline of each resource (i.e., ...). The categorized requests are then passed to the dependency constraint module.

[0076] The dependency constraint module receives classification request information from the resource collection management module and performs the following operations: (1) Playback synchronization constraint check, that is, for video segments and audio segments, the constraint module ensures that the audio and video are synchronized. For example, the download completion time of a video segment cannot be later than its playback deadline, otherwise it will cause playback stuttering.

[0077] (2) The latest start time is calculated according to the above description, and will not be repeated here.

[0078] (3) Quality of Experience (QoE) health monitoring relies on the constraint module to continuously monitor the amount of data in the player buffer. If the amount of data in the buffer is lower than the preset safety threshold, the obfuscation scheduling is suspended, video download is prioritized, and this status is notified to the hybrid orchestration module and the concurrent scheduling module.

[0079] Furthermore, the hybrid orchestration module performs resource splitting, dividing the video segment data into multiple video data blocks according to a set granularity (e.g., 100KB). Simultaneously, auxiliary resources (audio, images, subtitles, API responses) are also divided into auxiliary data blocks at the same granularity. Within the planning time window, with the optimization objective of maximizing the hybrid entropy of the output traffic sequence or minimizing the self-similarity of the output traffic sequence, an interleaved sequence is searched to obtain one that alternating the transmission order of video data blocks and auxiliary data blocks, with at least one auxiliary data block inserted between any two consecutive video data blocks. The generated interleaved sequence is output as a hybrid burst plan to the concurrent scheduling module for execution.

[0080] Multiple logical streams are initiated through a concurrent scheduling module, each assigned a different stream identifier (Stream ID). For example, a video data block is assigned Stream ID=1, an audio data block is assigned Stream ID=3, and an image data block is assigned Stream ID=5. Each data block is encapsulated into a data frame with the corresponding stream identifier and written to the underlying transmission buffer in a frame-level interleaved transmission sequence. The enqueue order of data frames with different Stream IDs in the transmission buffer is controlled with microsecond precision. When two adjacent data frames belong to different resource types and have different lengths, a padding frame is inserted between the two data frames to align the data frames from different resources on the transmission timeline.

[0081] After performing real-time calculation of the mixed entropy, the evaluation and feedback module performs a threshold judgment and adjusts the sequence based on the judgment result. After the adjustment is completed, the concurrent scheduling module continues to send data frames according to the updated sequence, and the evaluation and feedback module continues to monitor the changes in the mixed entropy, forming a continuous closed-loop feedback.

[0082] In other words, the collaboration between the above modules is as follows: the resource set management module intercepts and classifies resource requests, and passes the classification results to the dependency constraint module; the dependency constraint module calculates the latest start time and passes the constraints to the hybrid orchestration module; the hybrid orchestration module generates a frame-level interleaved transmission sequence based on the constraints and passes the sequence to the concurrent scheduling module; the concurrent scheduling module sends data frames according to the sequence and feeds back the transmission status to the evaluation feedback module; the evaluation feedback module calculates the hybrid entropy in real time, and when the hybrid entropy is lower than the threshold, it notifies the hybrid orchestration module to adjust the sequence, forming a closed-loop optimization.

[0083] The above modules enable deep interleaving of video stream data frames with data frames from other resources such as audio and images on the timeline without violating player buffer constraints. This makes it impossible for attackers to extract pure video features from the mixed traffic, thus effectively resisting traffic alignment analysis attacks and protecting user privacy and security.

[0084] According to another aspect of the embodiments of this application, a storage medium is also provided. Optionally, in this embodiment, the storage medium can be used to execute the program code of any of the above-mentioned anti-alignment download orchestration methods based on multi-resource hybrid bursting and multiplexing in the embodiments of this application.

[0085] Optionally, in this embodiment, the storage medium is configured to store program code for performing the following steps: S1, intercept the set of resource requests to be downloaded initiated by the client. The set of resource requests to be downloaded includes at least one target media stream resource request and at least one auxiliary resource request. The target media stream resource request is a video stream segmentation request. The auxiliary resource request includes an audio stream request, an image request, a subtitle request, or an interface response request. S2, based on the playback dependencies of each resource and the estimated bandwidth, calculate the latest start time corresponding to each resource request to ensure that the playback buffer is not lower than the safety level. S3, the video segment data corresponding to the target media stream resource request is divided into multiple video data blocks, and a frame-level interleaved transmission sequence is generated according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are alternately arranged or aggregated on the time axis to form a hybrid burst plan. S4. According to the frame-level interleaved transmission sequence, the corresponding request frame and data frame are sent to the server through different stream identifiers of the transmission protocol that supports multiplexing.

[0086] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated in this embodiment.

[0087] The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0088] According to another aspect of the embodiments of this application, an electronic device is also provided for implementing the above-described anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing, the electronic device being a server, a terminal, or a combination thereof.

[0089] Figure 4 This is a schematic diagram of the structure of an optional electronic device according to an embodiment of this application, such as... Figure 4 As shown, it includes a processor 402, a communication interface 404, a memory 406, and a communication bus 408. The processor 402, communication interface 404, and memory 406 communicate with each other via the communication bus 408. Memory 406 is used to store computer programs; When processor 402 executes a computer program stored in memory 406, it performs the following steps: S1, intercept the set of resource requests to be downloaded initiated by the client. The set of resource requests to be downloaded includes at least one target media stream resource request and at least one auxiliary resource request. The target media stream resource request is a video stream segmentation request. The auxiliary resource request includes an audio stream request, an image request, a subtitle request, or an interface response request. S2, based on the playback dependencies of each resource and the estimated bandwidth, calculate the latest start time corresponding to each resource request to ensure that the playback buffer is not lower than the safety level. S3, the video segment data corresponding to the target media stream resource request is divided into multiple video data blocks, and a frame-level interleaved transmission sequence is generated according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are alternately arranged or aggregated on the time axis to form a hybrid burst plan. S4. According to the frame-level interleaved transmission sequence, the corresponding request frame and data frame are sent to the server through different stream identifiers of the transmission protocol that supports multiplexing.

[0090] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated in this embodiment.

[0091] Optionally, the communication bus can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single thick line, but this does not indicate that there is only one bus or one type of bus. The communication interface is used for communication between the aforementioned electronic device and other devices.

[0092] Memory may include RAM or non-volatile memory. Volatile memory, for example, at least one disk storage device. Alternatively, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0093] The processors mentioned above can be general-purpose processors, including but not limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; they can also be DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0094] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0095] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0096] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0097] Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0098] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0099] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of embodiments of this disclosure upon considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

[0100] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0101] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for anti-alignment download orchestration based on multi-resource hybrid bursting and multiplexing, characterized in that, include: The client intercepts a set of resource requests to be downloaded, which includes at least one target media stream resource request and at least one auxiliary resource request. The target media stream resource request is a video stream segmentation request, and the auxiliary resource request includes an audio stream request, an image request, a subtitle request, or an interface response request. Based on the playback dependencies of each resource and the estimated bandwidth, calculate the latest start time corresponding to each resource request to ensure that the playback buffer is not lower than the safety level. The video segment data corresponding to the target media stream resource request is divided into multiple video data blocks, and a frame-level interleaved transmission sequence is generated according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are arranged alternately or aggregated on the time axis to form a hybrid burst plan. According to the frame-level interleaved transmission sequence, corresponding request frames and data frames are sent to the server through different stream identifiers of the transmission protocol that supports multiplexing.

2. The anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing as described in claim 1, characterized in that, The generation of the frame-level interleaved transmission sequence includes: Within the planning time window, with the optimization objective of maximizing the mixed entropy of the output traffic sequence or minimizing the self-similarity of the output traffic sequence, an interleaved sequence is searched to ensure that video data blocks and auxiliary data blocks are alternately arranged in the transmission order, and that at least one auxiliary data block is inserted between any two consecutive video data blocks. The planning time window is determined based on the latest start time and the current time. The mixed entropy is used to characterize the degree of mixing of data frames of different resource types in the output traffic sequence. The auxiliary data block is obtained by dividing the resource data corresponding to the auxiliary resource request according to a set granularity.

3. The anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing as described in claim 2, characterized in that, The method further includes: Calculate the mixed entropy of the current output flow sequence in real time; When the mixing entropy is lower than a preset mixing degree threshold, the frame-level interleaved transmission sequence is adjusted, including at least one of the following methods: delaying the transmission of some video data blocks, sending auxiliary data blocks in advance, changing the interleaving density of data blocks, or inserting camouflage data blocks.

4. The anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing as described in claim 2, characterized in that, The auxiliary data block is obtained by dividing the resource data corresponding to the auxiliary resource request according to a set granularity, specifically including: The set granularity of the video data block is obtained as the target granularity; The resource data corresponding to the auxiliary resource request is divided into multiple auxiliary data blocks, such that the size of each auxiliary data block matches the target granularity. The auxiliary resources include image resources, and the image resources are divided into multiple image data blocks corresponding to the number of video data blocks.

5. The anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing as described in claim 1, characterized in that, The method further includes: Monitor the current data volume in the player's buffer in real time; When the current data volume is lower than a preset security threshold, the execution of the hybrid burst plan is terminated. Prioritize scheduling and sending the video segment data corresponding to the target media stream resource request until the current data volume recovers to above the safety threshold, then resume execution of the hybrid burst plan.

6. The anti-alignment download orchestration method based on multi-resource hybrid bursting and multiplexing as described in claim 1, characterized in that, Sending corresponding request frames and data frames through different stream identifiers includes: Start multiple logical streams and assign different stream identifiers; According to the frame-level interleaved transmission sequence, each data block is encapsulated into a data frame with the corresponding stream identifier and written into the underlying transmission buffer. When two adjacent data frames belong to different resource types and have different lengths, a padding frame is inserted between the two data frames to align the data frames from different resources on the transmission timeline.

7. A multi-resource hybrid burst and multiplexing-based anti-alignment download orchestration device, characterized in that, include: The resource collection management module is used to intercept the set of resource requests to be downloaded initiated by the client. The set of resource requests to be downloaded includes at least one target media stream resource request and at least one auxiliary resource request. The target media stream resource request is a video stream segmentation request, and the auxiliary resource request includes an audio stream request, an image request, a subtitle request, or an interface response request. The dependency constraint module is used to calculate the latest start time corresponding to each resource request based on the playback dependency relationship of each resource and the estimated bandwidth, so as to meet the condition that the playback buffer is not lower than the safety level. The hybrid orchestration module is used to divide the video segment data corresponding to the target media stream resource request into multiple video data blocks, and generate a frame-level interleaved transmission sequence according to the latest start time. In the frame-level interleaved transmission sequence, the video data blocks and the auxiliary data blocks corresponding to the auxiliary resource request are alternately arranged or aggregated on the time axis to form a hybrid burst plan. The concurrent scheduling module is used to send corresponding request frames and data frames to the server according to the frame-level interleaved transmission sequence and through different stream identifiers of the transmission protocol that supports multiplexing.

8. The anti-alignment download orchestration device based on multi-resource hybrid bursting and multiplexing as described in claim 7, characterized in that, The device further includes: The evaluation feedback module is used to calculate the mixed entropy of the current output flow sequence in real time; When the mixing entropy is lower than a preset mixing degree threshold, the frame-level interleaved transmission sequence is adjusted to increase the mixing entropy of the current output traffic sequence.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program, when executed, performs the method of any one of claims 1 to 6.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to execute the method of any one of claims 1 to 6 through the computer program.

Images