Performance and Fault Management in Cellular Networks through Device-Network Cooperation

Participants

  • Georgia Tech:
  • Purdue:
    • PI: Saurabh Bagchi
    • GRAs: Nawanol Theera-Ampornpunt, Heng Zhang
  • AT&T Research
    • Kaustubh Joshi
    • Rajesh Panta

Funding

National Science Foundation

Project Scope

Cellular networks have become a part of the infrastructure that we rely on without pausing to think of the enormous complexity that underlies such networks. These networks serve an amazing diversity of mobile devices and in an amazing diversity of radio environments.

Failures in such networks are not uncommon and we take their consequences in our stride as a fact of life, grumbling about call drops and data disconnections. Similarly, we also encounter performance degradations, either in voice quality of calls or connection speeds. With the fast increasing amount of traffic on cellular networks and the cellular spectrum crunch, these conditions are expected to become more severe in the near to mid-term future. Further, in today’s networks, there is no facility to provide services to a device by orchestrating an entire network of geographically dispersed mobile devices, such as, for providing a “video” of the path of a tornado.

In this project, we put forward the view that rather than reactive management of failures or performance degradations, we need to build into cellular networks, proactive management of failures and performance degradations. Proactive management will involve predicting impending events of interest (such as, congestion or call drop) and then taking some mitigation action. The primary difficulty for performance or fault management as well as for facilitating orchestrated services is that, to the end devices, the network has been essentially a black box. Conditions about the network, such as, congestion or overload of some elements of the cellular infrastructure, are not made visible to the applications on the mobile end devices. Likewise, much of the application eco-system within a device is not made visible to the network. Thus, it is not known to the network what is the latency requirement of some application or what demand it is going to place in the near future on the network. Our solution approach will enable proactive fault and performance management through cooperation between the network and the devices. Our solution comprises the following steps.

  • Data analysis at the network elements and device. Cellular networks collect a wealth of data about individual mobile devices and aggregate quality of the network. This data is distributed among its various elements, such as, the Radio Network Controller or base stations. We will perform analysis of this data to determine events of interest that are likely to happen in the near future, such as, network congestion or disconnection of a device. Such data analysis done in our preliminary work has indicated that if one tries to do this without regard to the specific cell or the type of mobile device, then predictors are either inaccurate or imprecise. It has also indicated that the data analysis models have to evolve over time due to the dynamic nature of the environment. In our framework of cooperation between the network and devices, the data analysis will also use information from the individual devices, such as, the nature of traffic being generated by each application, for example, to time shift some delay tolerant traffic.
  • Mitigation actions. The prediction of an impending event of interest will trigger a root cause analysis. The root cause will map to mitigation actions, either initiated by the mobile device or by a network element. Examples of such mitigation actions are: the application on the mobile device indicating to the control system that some of its traffic is delay tolerant and such traffic can then be delayed with the intent of easing the congestion; the network forcing a handoff of the mobile device from one base station to another to pre-empt an impending disconnection.
  • Enhancing mobile services. A significant thrust of work will be new mobile services that can be enabled by network-device cooperation. We will explore how enhanced computation resources can be made available to the end device by offloading parts of computation to “cloudlets”, which will reside in edge network elements and the device will connect to different cloudlets as the device physically moves from one location to another. Distinct from all of the volume of work on computation offloading, this thrust will develop new techniques for handling dynamism and resource constraints of cellular network links and for coping with frequent mobility. Another aspect of this work will be to have a device enjoy functionality that is achieved through the collective — a large number of devices in a geographical region pooling together their resources, orchestrated by the network, for common good. A canonical example is determination of an oncoming tornado through putting together images from multiple phones. This is akin to “crowd sensing” but in this project, we will address significant new challenges arising from the need to handle a large number of devices, the short time scales, and the geospatial aspect of the information.

Current Project Thrusts

The current thrusts in the project are:

  1. Gaming is a popular activity on smartphones. However, smartphones have limited storage available. This, coupled with the growing size of mobile games, leads to the situa- tion where the user needs to limit the number of games installed on their phone at any one time. We are developing a novel technique for reducing the storage requirements of mobile games, so that the user can keep more applications installed simultaneously. The technique relies on predicting user’s data blocks and pre-fetching them from a cloud storage server with enough lookahead that the user’s quality of experience is not degraded.
  2. The rapid adoption of smartphones with different types of advanced sensors has led to an increasing trend in the usage of mobile crowdsensing applications. However, high energy consumption has been found to be detrimental to their wide-spread adoption. We propose a framework, resident in the cellular network, which orchestrates the sensing and communication at participating devices so as to achieve high energy efficiency. It also provides a simple programming abstraction and helper functions (like location estimation) to help the development of crowd-sensing applications. We show its benefit by applying it to an existing crowdsensing application and at scale through simulation using a real cellular network trace.

Publications

  • Habak, K., Ammar, M., Harras, K., Zegura, E., “FemtoClouds: Leveraging Mobile Devices to Provide Cloud Service at the Edge,” Proceedings of the 8th IEEE International Conference on Cloud Computing, June 2015, New York.
  • Mansy, A., Fayed, M., Ammar, M.,”Network-layer Fairness for Adaptive Video Streams,” Proceedings of the IFIP Networking Conference, Toulouse, France, May 2015.
  • Shi, C., Joshi, K., Panta, R., Ammar, M., Zegura, E., “CoAST: collaborative application-aware scheduling of last-mile cellular traffic,” MobiSys ’14 Proceedings of the 12th annual international conference on Mobile systems, applications, and services, 2014
  • Arshad, F. A., Maji, A., Mudgal, S., Bagchi, S., “Is Your Web Server Suffering from Undue Stress due to Duplicate Requests?.” Short Paper, At the 11th International Conference on Autonomic Computing (ICAC), pp. 105-111, June 18-20, 2014, Philadelphia, PA.
  • Tarun Mangla, Nawanol Theera-Ampornpunt, Mostafa Ammar, Ellen Zegura, Saurabh Bagchi, “Video Through a Crystal Ball: Effect of Bandwidth Prediction Quality on Adaptive Streaming in Mobile Environments,” Proceedings of Workshop on Mobile Video Delivery, 2016 (MoVid ’16).Tarun Mangla, Nawanol Theera-Ampornpunt, Mostafa Ammar, Ellen Zegura, Saurabh Bagchi, “Video Through a Crystal Ball: Effect of Bandwidth Prediction Quality on Adaptive Streaming in Mobile Environments,” Proceedings of Workshop on Mobile Video Delivery, 2016 (MoVid ’16).
  • Nawanol Theera-Ampornpunt, Tarun Mangla, Saurabh Bagchi, Rajesh Panta, Kaustubh Joshi, Mostafa Ammar, and Ellen Zegura, “Toward a More Reliable Mobile Streaming through Cooperation between Cellular Network and Mobile Devices,” In Proceedings of the IEEE 35th Symposium on Reliable Distributed Systems (SRDS), pp. 1-10, September 26-29, 2016, Budapest, Hungary.
  • Saurabh Bagchi, Nawanol Theera-Ampornpunt, Mostafa Ammar, Ellen Zegura, Tarun Mangla, Rajesh Panta, Kaustubh Joshi, “TANGO: Performance and Fault Management in Cellular Networks through Device-Network Cooperation,” In submission to IEEE Communications Magazine, September 2016.
Last modified: March 14, 2017