Search

EP-3981106-B1 - ALLOCATION OF FOG NODE RESOURCES

EP3981106B1EP 3981106 B1EP3981106 B1EP 3981106B1EP-3981106-B1

Inventors

  • M, Saravanan
  • BANERJEE, ARINDAM

Dates

Publication Date
20260513
Application Date
20190607

Claims (13)

  1. A method (300) for allocating resources of fog nodes, wherein the fog nodes are organised into at least one fog network; the method comprising: Receiving (332) a request from a client node; Identifying (334) requirements for fulfilment of the request; Determining (336) a location of the client node for fulfilment of the request, comprising estimating a trajectory of the client node on the basis of historical location data for the client node within an environment; Identifying (338), from the identified requirements and the determined location, a cluster of fog nodes operable to fulfil the request; and Selecting (340), from the identified cluster, fog nodes the resources of which are to be allocated to fulfil the request by minimising at least one of: a number of clusters required to fulfil the request; a total time to fulfil requests received by the fog nodes; a number of unfulfilled requests received by the fog nodes; wherein identifying (438), from the identified requirements and the determined location, a cluster of fog nodes operable to fulfil the request comprises: Generating (438a) at least one cluster comprising fog nodes that are: operable to at least partially fulfil the request; within a threshold distance of each other at a particular point in time; and operable to cooperate with one another.
  2. The method (300) as claimed in claim 1, wherein identifying (434) requirements for fulfilment of the request comprises: Identifying (434a) a primary service for the request, wherein the primary service comprises a service provided by at least one fog node, the service of which is necessary to fulfil the request; and Identifying (434b) fog nodes operable to provide the primary service for the request.
  3. The method (300) as claimed in any one of the preceding claims, wherein selecting (440), from the identified cluster, fog nodes the resources of which are to be allocated to fulfil the request comprises calculating an objective function minimising a weighted sum of terms representing at least one of: a number of clusters required to fulfil the request; a total time to fulfil requests received by the fog nodes; a number of unfulfilled requests received by the fog nodes.
  4. The method (300) as claimed in any one of the preceding claims, further comprising: Initiating (442) allocation of resources of the selected fog nodes to fulfil the received request.
  5. The method (300) as claimed in any one of the preceding claims, further comprising: Fulfilling (446) of the received request by the selected fog nodes.
  6. The method (300) as claimed in any one of the preceding claims, further comprising: Registering (402) a new fog node; and Associating (404) a service provided by the fog node with corresponding services provided by other fog nodes.
  7. The method (300) as claimed in any one of the preceding claims, further comprising: Registering (420) a new client node; Monitoring (422) a location of the registered client node; and Learning (424) from the monitored location at least one trajectory within an environment that is associated with the registered client node.
  8. The method (300) as claimed in any one of the preceding claims, further comprising: Modelling (426) an environment within which the fog nodes are located as a location graph comprising vertices representing waypoints within the environment and edges representing legitimate paths between the waypoints.
  9. The method (300) as claimed in claim 8, further comprising: Generating (428) a connectivity matrix of the modelled environment, the connectivity matrix representing connectivity between legitimate paths of the environment.
  10. The method (300) as claimed in claim 8 or 9, further comprising: Generating (428) a path type matrix of the modelled environment, the path type matrix representing an assigned type of the legitimate paths in the environment.
  11. The method (300) as claimed in any one of claims 8 to 10, further comprising: Generating (428) a waypoint dimension matrix of the modelled environment, the waypoint dimension matrix representing a number of legitimate paths through the environment that meet at each waypoint.
  12. A computer program (550) comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out a method (300) according to any one of the preceding claims.
  13. A controller (500) for allocating resources of fog nodes, wherein the fog nodes are organised into at least one fog network, the controller comprising a processor (502) and a memory (504), the memory containing instructions (550) executable by the processor (502) such that the controller (500) is operable to carry out a method (300) according to any one of claims 1 to 11.

Description

Technical Field The present disclosure relates to a method for allocating resources of fog nodes, wherein the fog nodes are organised into at least one fog network. The present disclosure also relates to a controller and to a computer program and a computer program product configured, when run on a computer to carry out a method for allocating resources of fog nodes. Background Fog computing refers to the extension of Cloud computing to the edge of a network, facilitating the operation of compute, storage, and networking services between end devices and cloud computing data centers. Fog computing may thus be considered as a complement to cloud computing, and is predicted to benefit varying domains including mobile/wearable computing, Internet of Things (IoT), and big data analytics. Some of the advantages afforded by fog computing include reducing latency, increasing throughput, consolidating resources, saving energy, and enhancing security and privacy. For example, in big data analytics, huge volumes of data are generated at the edge of network. Fog computing supports edge analytics, which can reduce the delay of big data analytics and decrease the cost of data transmission and storage. Fog computing is likely to be a key component in developing the smart city paradigm, which envisages the use of Information and Communication Technologies (ICT), to develop, deploy and promote sustainable development practices in order to address growing urbanisation challenges. Central to this ICT framework is an intelligent network of connected objects and machines that transmit data using wireless technology and the cloud. Cloud-based IoT applications receive, analyse and manage data in real time to help municipalities, enterprises and citizens make improved decisions. Traditional cloud computing architectures cannot meet the requirements of a massive IoT deployment on this scale, and Fog computing is therefore envisaged to address issues including latency, protection of network bandwidth, security concerns and reliability of operations, adaptation to different environmental conditions and the ability to both store data and take decisions closer to a relevant location. In some situations, the ability to analyse data close to the device that collected the data can make the difference between addressing an issue in time and a cascading system failure. The increasing cost of transport and speed of processing are also driving the adoption of fog computing, in which computing power is distributed across the edge network, data centers, and the public cloud. Fog computing is starting to be rolled out in smart cities, connected cars, drones and other applications. Key challenges in running IoT applications in a Fog computing environment are resource allocation and task scheduling. Fog computing research is still in its infancy, and taxonomy-based investigation into the requirements of Fog infrastructure, platform, and applications mapped to current research is still required. It will be appreciated that analysing IoT sensor data close to where it is collected has the advantages of minimising latency, offloading significant amounts of network traffic form the core network and maintaining potentially sensitive data within the network in which it was generated. Resources are most dynamic and heterogeneous in a Fog environment because of the diversity of IoT devices and their available resources. All devices known as Fog devices are responsible for performing the computation of their own application. Fog computing aims to use idle resources available on any Fog device, with Fog computation always taking second priority to the device's own application. In general, the amount of resources available for Fog computation is dynamic but predictable through analysis of long-term activity of Fog node resources. This prediction is useful because during execution of a Fog task, the status of resources of the Fog node carrying out the task may change, owing for example to receipt of a request from an application for which the Fog node is responsible. This differs from a Cloud situation, in which it is possible to know what resources are currently available and whether or not they are exclusively used for cloud-based application requests. Resource allocation and scheduling as a part of dynamic collaboration within Fog nodes are consequently considerably more challenging than resource allocation and scheduling in the cloud. Important characteristics of Fog computing such as low latency and location awareness, wide-spread geographical distribution, mobility, multiple nodes in close proximity, predominant role of wireless access, significant presence of streaming and real-time applications and heterogeneity. An example use case embodying some of these characteristic is a smart city situation involving four fog networks - smart building, smart transport, connected cars and smart parking. These networks are available in a smart city environment illu