System for merging sessions across multiple devices in over-the-top (OTT) media platforms

Abstract

A system for merging sessions from multiple devices in an over-the-top (OTT) media platform, wherein the system includes the following: a telemetry harmonization module configured to receive heterogeneous playback telemetry streams from a variety of client devices, wherein the telemetry harmonization module is further configured to normalize device-specific playback events, including pause, play, search, fast forward, rewind, buffer, and disconnect events, into a uniform scheme using a network-synchronized reference clock to mitigate device-side variations; a conflict resolution control unit configured to process normalized playback telemetry streams and detect simultaneous or conflicting playback events originating from different devices linked to the same user account; a stitching processing unit configured to create a canonical playback state by integrating the resolved playback telemetry streams, wherein the stitching processing unit applies a dynamic session stitching technique that prioritizes playback events based on at least recency, engagement intensity, and device hierarchy; and a canonical state repository configured to store and distribute the canonical playback state, wherein the repository comprises a distributed ledger structure anchored by cryptographic hashes to ensure the immutability and tamper-proof nature of the playback state across the multitude of client devices.

Description

Technical field

[0001] The present invention relates to the field of digital media streaming technologies and, in particular, to a system for harmonizing fragmented user sessions across multiple devices within over-the-top (OTT) media platforms. The invention also relates to computer devices and machine-implemented architectures configured to manage session continuity, intelligently merge playback states, and ensure seamless content delivery in heterogeneous user environments. Background of the invention

[0002] Over-the-top (OTT) platforms provide users with on-demand media content such as movies, TV series, and live events via the internet, bypassing traditional cable or satellite transmission. With the increasing prevalence of devices like smartphones, tablets, smart TVs, and connected streaming sticks, users frequently switch between devices during a media session. Traditional OTT platforms often struggle to accurately reconcile fragmented playback sessions because telemetry data is inconsistent, device status is not synchronized with central session data, and network conditions vary.

[0003] Existing solutions rely on simple "resume play" mechanisms that use only device-local playback timestamps, which are prone to discrepancies, inaccuracies, and inconsistencies. For example, if a user watches part of a movie on a smart TV and then continues watching on a mobile phone, the resume point may not accurately reflect the last access. This can be due to delays in telemetry data collection or conflicts arising from multiple devices simultaneously reporting different states. This leads to user dissatisfaction, increased churn, and lost monetization opportunities.

[0004] Therefore, there is a need for a technically robust system that can intelligently merge fragmented user sessions across multiple devices by harmonizing playback telemetry, applying conflict resolution techniques, and generating a canonical session state in real time.

[0005] The rise of over-the-top (OTT) media platforms has fundamentally changed how consumers access and consume digital content. Unlike traditional cable or satellite distribution systems based on a linear transmission model, OTT platforms deliver content over the internet, allowing users to watch what they want, when they want, and on any device. With the rapid proliferation of smart devices—from smartphones, laptops, tablets, and smart TVs to connected streaming dongles and game consoles—users are no longer tied to a single device for their media consumption. Instead, they seamlessly switch between devices depending on context, location, and convenience. A user might start watching a movie on the TV in the living room, continue watching it on a mobile device while commuting, and then pick up again later on a laptop.While this device-independent flexibility has become one of the most important drivers of OTT adoption, it also presents significant challenges in maintaining seamless user sessions across heterogeneous devices.

[0006] To ensure session continuity, OTT platforms initially implemented rudimentary mechanisms like the playback resume function, where a video's playback position was stored locally on the device. If a user paused playback and later resumed it on the same device, the video would continue from the last recorded point. This worked reasonably well in single-device scenarios but broke down completely when the same account was used across multiple devices. For example, if a user paused playback on a smart TV and then tried to resume it on a smartphone, the system often failed to synchronize the playback point. This resulted in discrepancies and forced the user to manually search for the approximate position.

[0007] The lack of centralized status synchronization rendered these early solutions inadequate for the multi-device ecosystem that has since become the norm.

[0008] With the evolution of OTT platforms, cloud-based session synchronization was introduced as an improvement. Platforms began storing playback progress and associated metadata in server-side databases instead of device-local caches. When a user switched devices, the new device queried the server to retrieve the most up-to-date playback status. While this method offered better consistency, it also introduced new problems. A major drawback was the latency in telemetry recording. Device-side playback events such as pause, fast-forward, or stop were often transmitted to the server asynchronously, resulting in delays or packet loss where the central record did not reflect the actual last viewed position. Network interruptions or offline playback further exacerbated the problem, leading to inaccurate resume points.

[0009] To improve accuracy, some platforms experimented with regular checkpoints. Instead of relying solely on start, pause, or stop signals, a device would report its playback position at fixed intervals. This allowed for a more detailed playback history but came with drawbacks. Frequent reports increased network bandwidth usage, led to higher server loads, and reduced device battery life in mobile environments. Furthermore, if a user switched devices between reporting intervals, the resume point could be several seconds or even minutes in the past, impacting the perceived seamlessness of the experience.

[0010] Another approach was to prioritize the "newest" event across devices. If two devices reported playback activity for the same account, the server prioritized the event with the most recent timestamp. While this seemed logical, it led to conflicts. Consider a scenario where a user paused playback on a smart TV but inadvertently left the app running in the background on a smartphone. The smartphone might continue sending telemetry data indicating minor playback activity, overwriting the intended TV session. The result was an incorrect resume point being recorded, causing the user to continue watching from an unintended location. This "last write wins" approach was too simplistic and often led to frustrating user experiences when multiple devices were logged in simultaneously.

[0011] To address this problem, some platforms implemented heuristic-based conflict resolution. These systems considered parameters such as device type, screen resolution, and playback duration to determine which device represented the "primary" session. For example, a smart TV session might be prioritized over a smartphone session, based on the assumption that users are more purposeful on larger screens. While this heuristic model reduced some erroneous overrides, it remained fragile and context-insensitive. Many scenarios existed where a user might genuinely prefer to continue playing on a mobile device—for instance, while traveling—and the heuristic miscalculation resulted in inaccurate state synchronization. Furthermore, heuristics were often hard-coded and inflexible, making them difficult to adapt to varying user behavior across different regions and demographics.

[0012] More sophisticated implementations attempted to create session-stitching processing units that aggregated telemetry data across devices and tried to reconcile it using probabilistic models. These models considered not only timestamps but also engagement signals such as playback duration, search activity, and interaction frequency. While this represented progress, most existing solutions struggled with scalability. OTT platforms serve millions of concurrent users, and telemetry data flows in massive amounts from a globally distributed device base. Resolving conflicts in real time using complex probabilistic techniques placed a significant strain on computing resources, leading to bottlenecks, delayed updates, and higher operating costs.

[0013] Another critical limitation of existing systems lies in their inability to guarantee data integrity and the immutability of replay states. In many implementations, replay states are stored in centralized databases, which are vulnerable to inconsistencies, rollbacks, or even manipulation in rare but damaging scenarios such as fraudulent activity or system compromise. Without cryptographic anchoring or distributed consensus, canonical replay states can diverge, leading to fragmented user experiences and complicating customer support or debugging.

[0014] Content protection mechanisms also complicate session stitching. Due to licensing restrictions, many OTT platforms implement Digital Rights Management (DRM) and secure streaming protocols. This creates additional hurdles for telemetry reporting, as devices often perform playback processes in a sandbox, limiting the fidelity of the data transmitted to the server. Existing systems based on device-side reporting struggle to obtain accurate and timely signals in DRM-restricted environments, further impacting session continuity.

[0015] Furthermore, the heterogeneity of devices presents a constant challenge. Different operating systems, player frameworks, and network stacks generate telemetry data in incompatible formats. Some devices report playback with absolute timestamps, others with relative offsets. Network latencies vary considerably depending on Wi-Fi, mobile data, and offline caching scenarios, making it difficult to achieve flawless cross-device synchronization of playback states. While standardization efforts such as the Common Media Application Format (CMAF) and MPEG-DASH ensure some consistency in content delivery, session state synchronization is not uniformly standardized, forcing platforms to develop their own proprietary and often unstable solutions.

[0016] Finally, user behavior itself introduces complexities that existing solutions inadequately address. Users often interact with multiple devices simultaneously, sometimes intentionally, sometimes unintentionally. For example, a family account might result in a child streaming on a tablet while a parent with the same profile watches on a smart TV, leading to playback conflicts. Existing systems lack sophisticated conflict resolution in such cases, resulting in erroneous overwrites and user dissatisfaction. Furthermore, users expect seamless continuity; that is, once they pause on one device, they can resume playback on another without any noticeable delay. Current architectures based on batch updates or regular queries fail to meet this expectation, leading to inefficiencies.

[0017] In summary, while numerous solutions have been developed to address the challenge of multi-device session continuity in OTT platforms, each has inherent drawbacks. Local storage approaches fail completely in cross-device scenarios. Server-side synchronization improves consistency but struggles with latency, bandwidth efficiency, and accuracy. Regular checkpointing consumes resources and still leaves gaps. Last-write-wins approaches oversimplify conflict resolution, while heuristic prioritization models lack adaptability. Probabilistic stitching units improve fidelity but often fail to scale efficiently to a global user base. Security, DRM restrictions, device heterogeneity, and the lack of standardized protocols further exacerbate the limitations of current methods.Consequently, no existing solution could provide reliable, seamless, accurate, and tamper-proof session stitching across heterogeneous devices in real time. This leaves a significant technological gap, which the present invention attempts to close. Summary of the invention

[0018] The invention provides a system for multi-device session stitching in over-the-top (OTT) media platforms, comprising a session stitching processing unit (SSE) implemented on a server-side computer device, a telemetry harmonization module (THM) configured to normalize playback events from heterogeneous devices, a conflict resolution control unit (CRE) for resolving timestamp discrepancies between sessions, and a canonical state repository (CSR) serving as a distributed ledger of active user sessions.

[0019] The system enables real-time capture of playback events, harmonizes fragmented telemetry data streams, and applies a stitching technique to create a unified playback state. The canonical session state is then redistributed to devices via a synchronization channel. This ensures that each device linked to the user account resumes playback from the most accurate point where it resumed.

[0020] The invention further discloses a machine / device embodiment in the form of a Session Stitching Appliance (SSA). This comprises a multi-core processor, high-speed memory, dedicated cryptographic accelerators for secure user mapping, and a machine-readable medium containing executable instructions for performing session stitching operations. The SSA can be deployed as a standalone appliance in a cloud-edge hybrid architecture or integrated into CDN edge nodes to enable low-latency session synchronization.

[0021] The main objective of the present invention is to provide a technically robust and scalable system for multi-device session stitching in over-the-top (OTT) media platforms, ensuring seamless continuity of the playback experience across heterogeneous devices without interruptions, latency, or loss of quality. The invention seeks to overcome the inherent limitations of existing resume-play mechanisms, server-based synchronization models, and heuristic conflict resolution approaches by introducing a more intelligent, harmonized, and tamper-proof framework for session continuity. A further objective of the invention is to establish a canonical playback state that dynamically integrates telemetry data streams from multiple devices while eliminating conflicts, discrepancies, and timestamp deviations.This ensures that a user can switch devices at any time and resume media content from the most accurate resume point.

[0022] A further objective of the invention is to provide an architectural framework that harmonizes fragmented telemetry data streams from different devices, operating systems, and playback systems and normalizes them into a unified scheme suitable for session synchronization. This harmonization ensures interoperability between the diverse devices that consumers use for media consumption, including smartphones, smart TVs, tablets, and networked streaming devices, thereby reducing fragmentation caused by heterogeneous device environments. The invention also aims to optimize resource utilization by developing stitching techniques and telemetry processing pipelines that minimize network bandwidth consumption, reduce server load, and enable low-latency synchronization even in large-scale implementations with millions of concurrent users.

[0023] Another important objective of the invention is to provide a conflict resolution mechanism that goes beyond simplified last-write-wins principles or rigid heuristic models. Instead, a multi-factorial decision unit is introduced that evaluates the intensity of interaction, the recency of events, and the contextual device hierarchy to determine the true canonical playback state. This ensures that erroneous overwrites caused by background telemetry or concurrent device sessions are minimized, thereby reducing user dissatisfaction. Furthermore, the invention aims to guarantee the cryptographic integrity and tamper-proof nature of session states by using blockchain anchoring or distributed ledger mechanisms to store canonical playback data.This ensures the immutability of the data, avoids rollback inconsistencies, and increases trustworthiness in session continuity processes.

[0024] A further objective of the invention is the integration of the system into a dedicated machine implementation in the form of a Session Stitching Appliance (SSA). This consolidates the computing, storage, and encryption functions required for executing the stitching operations. By deploying the SSA either in cloud infrastructures or at edge nodes of Content Delivery Networks (CDNs), stitching latency is to be minimized, and reliable, infrastructure-compatible continuity support for OTT platforms is to be provided. The invention also provides for the provision of a machine-readable medium containing executable instructions for stitching operations to ensure compatibility with various deployment scenarios.

[0025] Another objective of the invention is to improve the user experience by ensuring seamless transitions between devices with near-instantaneous playback resumption. This eliminates the need to manually search for or reload media streams. The invention also aims to improve platform-specific metrics such as playback duration, engagement, and retention by ensuring that users are not deterred by playback inconsistencies. From a commercial perspective, the invention also aims to secure revenue streams by minimizing user churn and more accurately tracking session behavior, which can lead to targeted recommendations and personalized content delivery strategies.

[0026] The overarching goal of the invention is ultimately to provide a comprehensive, technically advanced and scalable solution that addresses the long-standing problem of fragmented playback continuity in OTT environments, while ensuring the accuracy, integrity, adaptability and security of session composition across multiple devices in a way that has not been achieved with existing solutions. BRIEF DESCRIPTION OF THE FIGURE

[0027] These and other features, aspects, and advantages of the present invention will be better understood if the following detailed description is read with reference to the accompanying drawing, in which the same symbols consistently represent the same parts. The following applies: Fig. Figure 1 shows a block diagram of a system for multi-device session stitching in over-the-top (OTT) media platforms.

[0028] Experts will also recognize that the elements in the drawing are shown for the sake of simplicity and are not necessarily to scale. For example, the flowcharts illustrate the process by highlighting the main steps to enhance understanding of the aspects of this disclosure. Furthermore, with regard to the design of the device, one or more components of the device may be represented in the drawing by conventional symbols, and the drawing may show only the specific details relevant to understanding the embodiments of this disclosure, so as not to clutter the drawing with details that are readily apparent to those skilled in the art after reading this description. Detailed description of the invention

[0029] For a better understanding of the inventive principles, reference is made below to the embodiment shown in the drawing, which is described in specific terminology. However, this does not limit the scope of the invention. Changes and further modifications of the illustrated system, as well as further applications of the inventive principles, are possible, as would normally occur to a person skilled in the art in the field of invention.

[0030] It is clear to the person skilled in the art that the preceding general description and the following detailed description are exemplary and explanatory of the invention and are not intended as a limitation of it.

[0031] References in this specification to “an aspect”, “another aspect”, or similar expressions mean that a particular feature, structure, or property described in connection with the embodiment is included in at least one embodiment of the present disclosure. Therefore, occurrences of the expressions “in one embodiment”, “in another embodiment”, and similar expressions in this specification may all refer to the same embodiment, but need not.

[0032] The terms "includes," "include," or other variations thereof are intended to cover non-exclusive inclusion, such that a process or method that includes a list of steps may not only contain those steps but may also include other steps not expressly listed or inherent in such process or method. Likewise, the statement "includes..." in the case of one or more devices, subsystems, elements, structures, or components does not, without further limitations, preclude the existence of other devices, subsystems, elements, structures, components, or additional devices, subsystems, elements, structures, or components.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by a person skilled in the art in the field of the invention. The system, methods, and examples provided here serve only for illustration and are not to be construed as a limitation.

[0034] Embodiments of the present disclosure are described in detail below with reference to the attached drawing.

[0035] In Fig.Figure 100 is a block diagram of a system for multi-device session stitching in over-the-top (OTT) media platforms. The system comprises: a telemetry harmonization module (102) configured to receive heterogeneous playback telemetry streams from multiple client devices, the telemetry harmonization module also being configured to normalize device-specific playback events, such as pause, play, search, fast forward, rewind, buffer, and disconnect events, into a uniform scheme using a network-synchronized reference clock to minimize device-side variations; a conflict resolution control unit (104) configured to process the normalized playback telemetry streams and to detect concurrent or conflicting playback events from different devices associated with the same user account;a stitching processing unit (106) configured to generate a canonical playback state by integrating the resolved playback telemetry data streams, wherein the stitching processing unit applies a dynamic session stitching technique (106a) that prioritizes playback events at least according to recency, interaction intensity, and device hierarchy; and a canonical state repository (108) configured to store and distribute the canonical playback state, wherein the repository comprises a distributed ledger structure anchored by cryptographic hashes to ensure the immutability and tamper-proof nature of the playback state across the multitude of client devices.

[0036] In one embodiment, the telemetry harmonization module (102) also includes a timestamp correction subsystem that uses a hybrid synchronization model combining Network Time Protocol (NTP) offset correction and monotone event markers issued by the server, so that telemetry events from devices operating in different local time zones or experiencing time discrepancies are realigned to a single coherent time axis before being received into the conflict resolution control unit.

[0037] In one embodiment, the conflict resolution control unit (104) is also configured to perform a multi-factor decision function, wherein the function calculates a weighted priority rating for each playback event, using parameters that include: (a) Event timeliness based on corrected timestamps; (b) Engagement intensity, calculated as a continuous observation duration value, normalized over a sliding time window; (c) Weighting of the device hierarchy, whereby a session on a large-screen device conditionally takes precedence over a session on a handheld device, unless overridden by persistent interaction signals on the handheld device; and (d) Interaction density, measured as the frequency of active controls such as search or pause commands per unit of time. The decision function selects the playback event with the highest total score for inclusion in the canonical playback state.

[0038] In one embodiment, the canonical state repository (108) comprises a blockchain-anchored distributed ledger, wherein each update of the canonical replay state is recorded as a transaction, which includes the following: (a) a cryptographic hash of the normalized playback event stream that generated the update; (b) a digital signature that verifies the update as issued by the stitching processing unit; and (c) a pointer reference to a previous update of the canonical state, thereby forming a cryptographically linked sequence of canonical playback states that prevents rollback and provides verifiable verifiability of playback continuity.

[0039] In one embodiment, the stitching processing unit (106) also includes a synchronization distributor configured to distribute the canonical playback state in real time to all active client devices associated with a user account. The synchronization distributor uses low-latency communication protocols, including QUIC or WebSockets, and also employs delta encoding to transmit only incremental state changes, thereby reducing network overhead.

[0040] In one embodiment, the dynamic session stitching technique (106a) implemented by the stitching processing unit comprises a two-phase process, wherein a first phase includes a probabilistic inference of playback continuity across fragmented sessions by applying Bayesian sequence alignment to overlapping playback events, and a second phase includes a deterministic correction of the inferred states by enforcing logical playback consistency constraints, including a monotonic forward evolution of the playback position and the prevention of concurrent overlapping sessions within the same account.

[0041] In one embodiment, the telemetry harmonization module (102) also includes a DRM-enabled telemetry extraction subsystem configured for operation in sandbox playback environments. The subsystem receives telemetry data via secure APIs provided by the DRM framework, encrypts the telemetry with a platform-specific key, and transmits the data to the harmonization module without violating content protection restrictions.

[0042] In one embodiment, the canonical state repository (108) and the stitching processing unit (106) are deployed on a session stitching appliance at edge nodes of a content delivery network (CDN). The appliance includes a multi-core processing unit, high-speed non-volatile memory, cryptographic accelerators for blockchain anchoring, and low-latency network adapters, thereby reducing the round-trip latency for session voting and enabling playback to resume on multiple devices in less than one second.

[0043] In one embodiment, it further comprises a non-volatile, machine-readable medium on which instructions are stored which, when executed by a processor, cause the processor to perform operations that include the following: (a) Recording playback telemetry streams from heterogeneous devices; (b) Applying a timestamp correction to normalize telemetry; (c) Executing the multi-stage conflict resolution technique for selecting canonical playback events;( (d) Creating and storing a cryptographically anchored canonical playback state; and (e) Synchronizing the canonical playback status with all active client devices in real time.

[0044] In one embodiment, the conflict resolution control unit (104) is also configured to handle concurrent device playback scenarios by dynamically splitting user profiles into virtual subsessions when overlapping playback sessions are detected, with each subsession being assigned its own canonical playback thread, thereby preventing cross-device overrides while maintaining consistency within each concurrent playback stream.

[0045] The system is structurally enabled by specific hardware and tightly integrated, software-accelerated subsystems: The telemetry harmonization module consists of network interface controllers, API gateway devices, and edge ingestion devices with hardware time stamping (IEEE 1588 PTP or GPS-based real-time clocks) and dedicated I / O controllers that accept heterogeneous replay telemetry, perform device-specific parsing on FPGA- or CPU-based parsers, and normalize events into a uniform scheme, while data is provided in high-throughput NVMe caches to maintain order and minimize device-side variations;The conflict resolution control unit is physically implemented on low-latency compute nodes or FPGA-accelerated pattern detectors connected by a high-speed fabric, where stream processing engines (running on CPU / GPU or FPGA) correlate normalized events, manage per-account event queues in RAM-backed storage, and execute rule engines that detect concurrent or conflicting playback events using window comparisons and sequence number checks;The stitching processing unit is embodied as a series of containerized microservices running on multi-core servers or GPU / TPU-equipped nodes that implement the dynamic session stitching algorithm, including modules for freshness assessment, engagement intensity calculation (derived from event metadata and buffered telemetry), and device hierarchy resolution, with model inference offloaded to accelerators and canonical intermediate states maintained in key-value stores in main memory for atomic merging operations;The canonical state repository is implemented as a distributed storage fabric consisting of validation nodes or ledger devices with append-only NVMe journals, cryptographic hash engines, and hardware security modules for signing, as well as a replicated ledger structure (based on blockchain or Merkle tree) that ensures immutability and tamper resistance, while secure replication protocols, consensus controllers, and authenticated APIs distribute the canonical state to client devices. The concrete connection of network cards, timestamping hardware, accelerators, secure enclaves, high-throughput storage, and network structures enables a concrete, practical activation of each of the aforementioned elements and of the system as a whole.

[0046] The invention relates to a system for multi-device session stitching in over-the-top (OTT) media platforms. This system harmonizes, resolves, and consolidates fragmented playback telemetry data from multiple heterogeneous client devices into a canonical playback state that can be redistributed to all associated devices with cryptographic integrity and real-time synchronization. The architecture is designed to overcome fundamental limitations of existing solutions by integrating precise timestamp normalization, intelligent conflict resolution, and cryptographically anchored session persistence—all within a scalable system that can be operated in cloud and edge environments.

[0047] The system is based on the coordinated function of four primary subsystems: a telemetry harmonization module, a conflict resolution control unit, a stitching processing unit, and a canonical state repository. The telemetry harmonization module acts as the first point of contact for playback events from client devices. Each device—smartphones, tablets, smart TVs, set-top boxes, or streaming devices—generates playback telemetry in device-specific formats. These telemetry events include individual actions such as play, pause, search, rewind, fast forward, buffer, and disconnect. Due to the diversity of device platforms and network environments, such events are timestamped using local system clocks, which often differ due to time zone variations, clock errors, or temporary offline activity.

[0048] To ensure accurate session compilation, the telemetry harmonization module uses a timestamp correction subsystem that employs a hybrid synchronization technique. This technique combines standard Network Time Protocol (NTP) corrections with server-generated monotonic event markers. As soon as a device reports playback telemetry, its local timestamp is aligned with a server-issued monotonic marker to create a corrected event timeline. This alignment ensures that telemetry data from disparate devices is harmonized into a coherent temporal sequence before being received by the conflict resolution control unit.

[0049] After harmonization, the telemetry events are forwarded to the conflict resolution control unit, which is responsible for detecting and resolving playback conflicts. Conflicts arise when multiple devices linked to the same account generate overlapping or conflicting playback signals. For example, a smart TV session might display a pause event at a given time, while a smartphone logged into the same account simultaneously sends a playback event indicating continued use. A simplified "last-write-wins" approach would cause one of these events to overwrite the other without context, potentially resulting in an incorrect playback status.

[0050] To avoid this, the conflict resolution control unit executes a multi-stage decision function that calculates a weighted priority value for each event. This value is derived from several parameters: the event's recency, determined by corrected timestamps; the interaction intensity, measured by the duration of uninterrupted playback; the device hierarchy, which prioritizes sessions on large-screen devices such as televisions unless overridden by continued interaction on handheld devices; and the interaction density, which reflects the number of active user interactions such as searches or pauses. The conflict resolution control unit then selects the playback event with the highest overall value as the canonical input for the stitching processing unit.This multidimensional evaluation avoids incorrect overwrites by background applications and ensures that the user intent is captured faithfully in the canonical state.

[0051] The stitching processing unit implements the central session stitching technique, which operates in two phases. In the first phase, the unit applies probabilistic inference using Bayesian sequence alignment to correlate fragmented playback telemetry from different devices. The technique treats each event stream as a sequence and attempts to align it by calculating the posterior probabilities of continuity at each overlapping segment. For example, if a user pauses on one device after 42 minutes and resumes on another device after approximately 42 minutes and 15 seconds, the Bayesian model infers with a high probability that both refer to the same logical continuation of playback. In the second phase, the technique applies deterministic consistency constraints to ensure a logical playback progression.These constraints include a monotonous forward evolution of the playback position, ensuring that derived states never extend beyond previously defined canonical points, and the prevention of simultaneous overlaps of canonical sessions within a single profile. The combination of probabilistic inference and deterministic correction guarantees both flexibility in handling incomplete telemetry and accuracy in maintaining logical session continuity.

[0052] Once the stitching technique generates a canonical replay state, it is stored in the canonical state repository. This repository is implemented as a blockchain-based distributed ledger, with each update recorded as a transaction linked to the previous one via a cryptographic hash. Each transaction contains a hash of the normalized replay event stream that generated the update, a digital signature from the stitching processing unit to confirm authenticity, and a pointer to the immediately preceding canonical state. This architecture ensures the immutability of the session history, prevents rollback inconsistencies, and provides an auditable, tamper-proof record of replay continuity.

[0053] To ensure a seamless user experience, the canonical playback state must be distributed to all active client devices in real time. The stitching processing unit therefore includes a synchronization distributor that sends incremental state updates to the devices via low-latency communication protocols such as QUIC or WebSockets. To reduce bandwidth overhead, the distributor transmits updates using delta encoding, passing only incremental changes to the playback state instead of retransmitting the entire state. Devices receiving the update adjust their local playback pointers accordingly. This ensures that playback resumes at the correct canonical position without any noticeable delay when switching from one device to another.

[0054] For scenarios involving Digital Rights Management (DRM), special provisions are in place. In such cases, direct access to playback events is often restricted by sandbox execution environments. To address this issue, the telemetry harmonization module includes a DRM-enabled telemetry extraction subsystem that interacts with secure APIs from DRM frameworks. This subsystem encrypts telemetry data with platform-specific keys before transmitting it to the server. This ensures both compliance with content protection requirements and the availability of precise playback signals for stitching.

[0055] Scalability and latency optimization are achieved by deploying the stitching processing unit and canonical state repository on a dedicated appliance called the Session Stitching Appliance (SSA). The SSA comprises a multi-core processing unit coupled with GPU accelerators optimized for real-time event alignment, fast non-volatile memory for caching active telemetry data streams, and cryptographic accelerators for blockchain anchoring. The appliance is also equipped with low-latency network adapters to handle high-throughput communication with client devices.By using SSAs at the edge nodes of the Content Delivery Network (CDN), round-trip latency for session coordination is minimized, enabling playback status synchronization at intervals of less than one second, even during large streaming events with millions of concurrent users.

[0056] In scenarios where simultaneous playback on multiple devices is intended, such as when several family members use the same profile at the same time, the conflict resolution control unit dynamically partitions the profile into virtual subsessions. Each subsession is assigned an independent canonical thread, ensuring internal consistency while preventing cross-overrides. This guarantees a seamless experience for all concurrent viewers without interference from other viewers' telemetry.

[0057] The invention further provides for the provision of a non-volatile, machine-readable medium on which instructions are stored that, when executed by a processor, perform the described operations. These instructions direct the processor to acquire playback telemetry data, perform timestamp corrections, execute the multi-factor conflict resolution technique, generate cryptographically anchored canonical playback states, and distribute incremental updates to client devices.

[0058] By combining these modules and techniques, the invention offers a robust solution for multi-device session stitching in OTT platforms. Timestamp harmonization ensures temporal accuracy, the multi-factor conflict resolution ensures context-dependent fidelity, the probabilistic-deterministic stitching processing unit guarantees logical session continuity, and the blockchain-based repository ensures tamper resistance. The session stitching appliance provides the infrastructural capability to scale these processes globally with low latency, thus ensuring seamless, precise, and secure playback continuity that surpasses the limitations of existing solutions.

[0059] The present invention relates generally to digital media streaming technologies and session management architectures within over-the-top (OTT) platforms. More specifically, the invention relates to a system and a device for harmonizing fragmented playback telemetry across heterogeneous client devices, resolving conflicts when using multiple devices, and generating a canonical playback state that ensures seamless continuation of content across all devices. The invention lies at the intersection of distributed computing, cryptographic integrity assurance, real-time data synchronization, and media session stitching techniques, and is particularly relevant for large-scale OTT services that require robust, precise, and low-latency session continuity across a global user base.

[0060] The drawing and the preceding description show examples of embodiments. Those skilled in the art will recognize that one or more of the described elements can be combined to form a single functional element. Alternatively, certain elements can be divided into several functional elements. Elements of one embodiment can be added to another embodiment. For example, the sequence of the processes described here can be changed and is not limited to the manner described here. Furthermore, the actions of a flowchart need not be implemented in the sequence shown; nor does it necessarily have to be performed by all actions. Actions that are not dependent on other actions can also be performed in parallel with the other actions. The scope of the embodiments is in no way limited by these specific examples.Numerous variations are possible, whether explicitly stated in the specification or not, such as differences in structure, dimensions, and material usage. The range of embodiments is at least as broad as specified in the following claims.

[0061] Advantages, further benefits, and problem solutions have been described above with reference to specific embodiments. However, the advantages, benefits, problem solutions, and all components that can lead to an advantage, benefit, or solution occurring or becoming more apparent are not to be construed as critical, necessary, or essential features or components of individual or all claims. REFERENCES 100 A System For Multi-Device Session Stitching In Over-The-Top (OTT) Media Platforms. 102 Telemetry Harmonization Module 104 Conflict Resolution Control Unit 106 stapling processing units 108 Kanonisches Status- Repository

Claims

[1] A system for merging sessions from multiple devices in an over-the-top (OTT) media platform, wherein the system comprises: a telemetry harmonization module configured to receive heterogeneous playback telemetry streams from a variety of client devices, wherein the telemetry harmonization module is further configured to normalize device-specific playback events, including pause, play, search, fast forward, rewind, buffer, and disconnect events, into a uniform scheme using a network-synchronized reference clock to mitigate device-side variations; a conflict resolution control unit configured to process normalized playback telemetry streams and detect simultaneous or conflicting playback events originating from different devices linked to the same user account; a stitching processing unit configured to create a canonical playback state by integrating the resolved playback telemetry streams, wherein the stitching processing unit applies a dynamic session stitching technique that prioritizes playback events based on at least recency, engagement intensity, and device hierarchy; and a canonical state repository configured to store and distribute the canonical playback state, wherein the repository comprises a distributed ledger structure anchored by cryptographic hashes to ensure the immutability and tamper-proof nature of the playback state across the multitude of client devices. [2] System according to claim 1, wherein the conflict resolution control unit is further configured to execute a multi-stage decision function, the function calculating a weighted priority score for each playback event using the following parameters: (a) event freshness based on corrected timestamps; (b) interaction intensity, calculated as a continuous playback duration value normalized over a sliding time window; (c) device hierarchy weighting, wherein a session on a large-screen device is conditionally prioritized over a session on a handheld device unless overridden by sustained interaction signals on the handheld device; and (d) interaction density, measured as the frequency of active controls such as search or pause commands per unit of time, the decision function thereby selecting the playback event with the highest overall score for inclusion in the canonical playback state. [3] System according to claim 1, wherein the stitching processing unit further comprises a synchronization distributor configured to distribute the canonical playback status in real time to all active client devices associated with a user account, wherein the synchronization distributor uses low-latency communication protocols and also applies delta encoding to transmit only incremental status changes, thereby reducing network overhead. [4] System according to claim 1, wherein the dynamic session stitching technique implemented by the stitching processing unit comprises a two-phase process, wherein a first phase comprises a probabilistic inference of playback continuity across fragmented sessions by applying a Bayesian sequence alignment to overlapping playback events, and a second phase comprises a deterministic correction of the derived states by enforcing logical playback consistency constraints, including a monotonic forward evolution of the playback position and the prevention of simultaneous overlapping sessions within the same account. [5] System according to claim 1, wherein the telemetry harmonization module further comprises a DRM-enabled telemetry extraction subsystem configured for operation in sandbox playback environments, wherein the subsystem receives telemetry data via secure APIs provided by the DRM framework, encrypts the telemetry with a platform-specific key and transmits the data to the harmonization module without infringing the content protection restrictions. [6] System according to claim 1, wherein the canonical state repository and the stitching processing unit are deployed on a session stitching device at edge nodes of a Content Delivery Network (CDN), the device comprising a multi-core processing unit, non-volatile high-speed memory, cryptographic accelerators for blockchain anchoring, and low-latency network adapters, thereby reducing the round-trip latency for session voting and enabling resumption of playback on multiple devices in less than one second. [7] System according to claim 1, wherein the conflict resolution control unit is further configured to handle scenarios with simultaneous device playback by dynamically splitting user profiles into virtual subsessions when overlapping playback sessions are detected, with each subsession being assigned its own canonical playback thread, thereby preventing cross-device overrides while maintaining consistency within each simultaneous playback stream.

Images