System, device, and method of differentiating between streaming live-video flows and streaming non-live-video flows

A computerized method comprising:

• (a) deploying a Live-Video/Non-Live-Video detector unit,
• connected and operable between a core cellular network and an entry node of the Internet;
• (b) monitoring data packets exchanged over said core cellular network and over the Internet between a User Equipment (UE) device and a destination device;
• (c) determining that a particular communication flow between the UE device and the destination device is a streaming video communication flow;
• (d) further determining whether said streaming video communication flow is either (i) a streaming Live-Video communication flow or (ii) a streaming Non-Live-Video communication flow;
• (e) triggering deployment of one or more pre-defined operations if it was determined in step (d) that said streaming video communication flow is a streaming Live-Video communication flow and not a streaming Non-Live-Video communication flow;
• wherein the method comprises:
• selecting and applying a particular Machine Learning (ML) model which utilizes at least the following feature:
• a feature that (a) defines a window of N consecutive outgoing packets that are outgoing from the UE device and that are associated with a same single streaming video, and (b) performs a check whether at least M outgoing packets, out of said N consecutive outgoing packets, wherein M<N, have individually a byte-size that is greater than a first threshold value V1 and is smaller than a second threshold value V2;
• and (c) reaches a Live Video determination if the check result is positive, and reaches a Non Live Video determination if the check result is negative;
• wherein the threshold values V1 and V2 are pre-defined based on an identity of a source server that was requested by the UE device to stream said single streaming video to said UE device;
• wherein the method is implemented by at least a hardware processor.

A non-transitory storage medium having stored thereon instructions that, when executed by one or more hardware processors, cause the one or more hardware processors to perform a method comprising:

• (a) deploying a Live-Video/Non-Live-Video detector unit,
• connected and operable between a core cellular network and an entry node of the Internet;
• (b) monitoring data packets exchanged over said core cellular network and over the Internet between a User Equipment (UE) device and a destination device;
• (c) determining that a particular communication flow between the UE device and the destination device is a streaming video communication flow;
• (d) further determining whether said streaming video communication flow is either (i) a streaming Live-Video communication flow or (ii) a streaming Non-Live-Video communication flow;
• (e) triggering deployment of one or more pre-defined operations if it was determined in step (d) that said streaming video communication flow is a streaming Live-Video communication flow and not a streaming Non-Live-Video communication flow;
• wherein the determining of step (d) comprises:
• selecting and applying a particular Machine Learning (ML) model which utilizes at least the following feature:
• a feature that (a) defines a window of N consecutive outgoing packets that are outgoing from the UE device and that are associated with a same single streaming video, and (b) performs a check whether at least M outgoing packets, out of said N consecutive outgoing packets, wherein M<N, have individually a byte-size that is greater than a first threshold value V1 and is smaller than a second threshold value V2;
• and (c) reaches a Live Video determination if the check result is positive, and reaches a Non Live Video determination if the check result is negative;
• wherein the threshold values V1 and V2 are pre-defined based on an identity of a source server that was requested by the UE device to stream said single streaming video to said UE device.