Xapp conflict mitigation framework
According to an aspect, there is provided an apparatus configured to perform the following. The apparatus hosts a plurality of applications for managing network functions of a radio access network and schedules them for one or more future time slots. The apparatus obtains one or more predicted future values of one or more network parameters of the radio access network for the one or more future time slots. The apparatus determines whether or not at least one conflict exists in the scheduling of the plurality of applications based on the one or more predicted future values of the one or more network parameters. In response to determining that the at least one conflict exists, the apparatus resolves the at least one conflict by adjusting scheduling of at least one application involved in the at least one conflict.
1 . An apparatus comprising:
at least one processor; and
at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform:
hosting a plurality of applications for managing network functions of a radio access network;
scheduling the plurality of applications for one or more future time slots;
obtaining one or more predicted future values of one or more network parameters of the radio access network for the one or more future time slots;
determining whether or not at least one conflict exists in the scheduling of the plurality of applications based on the one or more predicted future values of the one or more network parameters; and
in response to determining that the at least one conflict exists, resolving the at least one conflict by adjusting scheduling of at least one application involved in the at least one conflict.
2 . The apparatus of claim 1 , wherein the obtaining of the one or more predicted future values of the one or more network parameters of the radio access network comprises:
maintaining, in the at least one memory, a trained machine-learning model, wherein one or more features of the trained machine-learning model comprise one or more cell and/or user parameters measurable in the radio access network, and one or more labels of the trained machine-learning model comprise the one or more network parameters of the radio access network;
obtaining one or more measured values for the one or more cell and/or user parameters; and
predicting, using the trained machine-learning model, one or more future values of the one or more network parameters for the one or more future time slots based on at least the one or more measured values of the one or more cell and/or user parameters.
3 . The apparatus according to claim 2 , wherein the one or more cell and/or user parameters comprise, for each of one or more terminal devices, at least parameters for defining a location of a terminal device within its serving cell and a most probable next serving cell of the terminal device.
4 . The apparatus of claim 2 , wherein the one or more cell and/or user parameters comprise
a cell area of a serving cell
a speed of one or more terminal devices and
at least one of
an angle of arrival from one or more terminal devices,
an angle of departure for transmission to one or more terminal device,
a beam identifier for a beam for transmission to or reception from one or more terminal devices
an antenna tilt for transmission to or reception from one or more terminal devices or
an antenna beam direction for transmission to or reception from one or more terminal devices.
5 . The apparatus according to claim 2 , wherein the apparatus is a near-real-time radio access network intelligent controller, near-RT RIC, of an open radio access network, O-RAN, architecture, and the plurality of applications are xApps,
the at least one memory and the instructions being configured, with the at least one processor, to cause the apparatus to perform, before the maintaining:
receiving the trained machine-learning model from a non-real-time radio access network intelligent controller, non-RT RIC, of the O-RAN architecture; and
storing the trained machine-learning model to the at least one memory.
6 . The apparatus according to claim 1 , wherein each or at least one of the at least one conflict corresponds to a conflict between
a value of at least one parameter defined by one of the plurality of applications based on the scheduling and the one or more predicted future values of the one or more network parameters and
a value of the at least one parameter defined by another of the plurality of applications based on the scheduling and the one or more predicted future values of the one or more network parameters,
wherein the conflict is associated with at least one of the one or more future time slots.
7 . The apparatus according to claim 1 , wherein each or at least one of the at least one conflict is one of:
a conflict in at least one key performance indicator to be affected by at least two applications,
a conflict in at least one network configuration parameter to be set by at least two applications,
a conflict due to a logical dependency between objectives of at least two applications,
a conflict due to an execution of at least one application using outdated measurement data and at least one other application using up-to-date measurement data and
a conflict in simultaneous allocation of radio resources by at least two applications.
8 . The apparatus according to claim 1 , wherein each or at least one of the at least one conflict is a conflict in scheduling of at least one application of the plurality of applications resulting from a change in a radio access network topology as predicted by the one or more predicted future values of the one or more network parameters of the radio access network.
9 . The apparatus according to claim 1 , wherein the adjusting of the scheduling of the at least one of the plurality of applications comprises:
suspending execution of the at least one application to avoid execution of the at least one application during at least one of the at least one future time slot.
10 . The apparatus according to claim 1 , wherein the at least one memory and the instructions are configured, with the at least one processor, to cause the apparatus to perform:
assigning a service level agreement violation penalty associated with a performance degradation to each or at least one of the at least one application whose scheduling is adjusted.
11 . The apparatus according to claim 1 , wherein the at least one memory and the instructions are configured, with the at least one processor, to cause the apparatus to perform:
preventing changing of values of one or more cell and/or user parameters associated with the at least one application for at least one future time slot for avoiding further conflicts resulting from the at least one conflict.
12 . The apparatus of claim 11 , wherein the at least one memory and the instructions are configured, with the at least one processor, to cause the apparatus to perform:
maintaining, in the at least one memory, information on one or more nested conflict scenarios where a conflict between the plurality of applications is known to result or potentially result in further conflicts later on; and
performing the preventing of the changing of values of the one or more cell and/or user parameters based on the information on the one or more nested conflict scenarios.
13 . The apparatus according to claim 1 , wherein the at least one memory and the instructions are configured, with the at least one processor, to cause the apparatus to perform:
maintaining, in the at least one memory, information on one or more known conflict scenarios involving any of the plurality of applications; and
performing the determining whether or not the at least one conflict exists in the scheduling of the plurality of applications further based on the one or more known conflict scenarios.
14 . The apparatus according to claim 1 , wherein the one or more network parameters of the radio access network comprise at least one of:
number of terminal devices in a serving cell,
number of terminal devices in one or more neighboring cells,
a quality of service of one or more terminal devices in the serving cell,
a quality of service of one or more terminal devices in one or more neighboring cells,
a downlink bit rate of one or more terminal devices in the serving cell,
an uplink bit rate of one or more terminal devices in the serving cell,
a cell individual offset for the serving cell,
a cell individual offset for one or more neighboring cells,
a hysteresis associated with one or more terminal devices in the serving cell,
a reference signal received power of one or more neighboring cell for one or more terminal devices,
a reference signal received power of one or more serving cells for one or more terminal devices,
a reference signal received quality of the one or more neighboring cells for one or more terminal devices,
a reference signal received quality of the one or more serving cells for one or more terminal devices,
a throughput of the serving cell,
a throughput of the one or more neighboring cells,
a signal-to-noise ratio for the serving cell,
a signal-to-noise ratio for the one or more neighboring cells,
a signal-to-interference-plus-noise ratio for the serving cell, or
a signal-to-interference-plus-noise ratio for the one or more neighboring cells.
15 . The apparatus according to claim 1 , wherein the plurality of applications are applications for providing value-added services, and the plurality of applications comprise at least one third-party application.
16 . The apparatus according to claim 1 , wherein the apparatus is a near-real-time radio access network intelligent controller, near-RT RIC, of an open radio access network, O-RAN, architecture, and the plurality of applications are xApps.
17 . A system comprising:
an apparatus of claim 16 ; and
a non-RT RIC comprising at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the non-RT RIC at least to perform:
maintaining, in the at least one memory, a training data set, wherein the training data set comprises a plurality of values of the one or more network parameters of the radio access network and a plurality of values of the one or more cell and/or user parameters of the radio access network;
initializing a machine-learning model;
training the machine-learning model using the training data set, wherein, in the training, the one or more cell and/or user parameters are features of the machine-learning model, and the one or more network parameters are labels of the machine learning model; and
transmitting the trained machine-learning model to the apparatus.
18 . A method comprising:
hosting a plurality of applications for managing network functions of a radio access network;
scheduling the plurality of applications for one or more future time slots;
obtaining one or more predicted future values of one or more network parameters of the radio access network for the one or more future time slots;
determining whether or not at least one conflict exists in the scheduling of the plurality of applications based on the one or more predicted future values of the one or more network parameters; and
in response to determining that the at least one conflict exists, resolving the at least one conflict by adjusting scheduling of at least one application involved in the at least one conflict.
19 . A non-transitory computer readable medium having stored thereon instructions that, when executed by a computing device, cause the computing device to perform:
hosting a plurality of applications for managing network functions of a radio access network;
scheduling the plurality of applications for one or more future time slots;
obtaining one or more predicted future values of one or more network parameters of the radio access network for the one or more future time slots;
determining whether or not at least one conflict exists in the scheduling of the plurality of applications based on the one or more predicted future values of the one or more network parameters; and
in response to determining that the at least one conflict exists, resolving the at least one conflict by adjusting scheduling of at least one application involved in the at least one conflict.