Publications

Controller Placement for Minimum Control Traffic in OpenDaylight Clustering

Authors: M. Karatisoglou, K. Choumas and T. Korakis

Conference:  IEEE WF-5G 2019, Dresden, Germany, 30 September - 02 October 2019

Abstract: Software-defined networking (SDN) decouples the control and data planes and moves the control logic to the SDN controllers. The controllers are servers handling the switches and directing their traffic with the use of open and non-proprietary protocols. Consequently, the controller is crucial for the network performance. OpenDaylight (ODL) is a widely recognised controller platform implementing a clustering mechanism, where multiple controllers are synchronized to behave as a single one, while also enabling load balancing and failure resilience. The purpose of this paper is to extract the communication patterns between the ODL controllers in a cluster, as well as between the controllers and the switches, in order to place the cluster controllers appropriately in a network. Our testbed experimentation reveals the bandwidth requirements of the control traffic in an ODL cluster, while a network model is introduced for the problem of determining the optimal controller placement for minimum control traffic.

Download paper: [Paper]

Enabling Distributed Spectral Awareness for Disaggregated 5G Ultra-Dense HetNets

Authors: K. Chounos, N. Makris and T. Korakis

Conference:  in proceedings of IEEE WF-5G 2019, Dresden, Germany, 30 September - 02 October 2019.

Abstract: Future wireless infrastructures will have to deal with significantly higher loads of data, produced by applications demanding higher capacity with lower latency for the exchanged information. 5G networks are expected to resolve these problems, by bringing in advanced flexibility of deployment through the Cloud-RAN concept, and new air interfaces. Dense Heterogeneous Networks will also assist in the smooth transition to a holistic approach for managing the wireless network, especially through their integration at the base station level. Nevertheless, networks operating under a very tight spectral environment, with limited spectrum resources, need to be aware of their expected performance in order to optimally decide their operating frequency. In this work, we consider a disaggregated heterogeneous base station, complying with the Cloud-RAN concept, and integrate spectral aware non-3GPP access (WiFi) to the base station.
We use multi-technology links to serve the end users, through two different paths: an LTE and a WiFi. The WiFi units are able to automatically discover their operating frequency ensuring efficient data delivery to the end users. We integrate all our contributions in a framework and deploy it over a real testbed. Our experimental results illustrate enhanced performance for the non-3GPP access of the network, delivering more than 5 times better throughput for high interference scenarios.

Download paper: [Paper]

Matching Theory Application for Efficient Allocation of Indivisible Testbed Resources

Authors: D. Stavropoulos, V. Miliotis, T. Korakis and L. Tassiulas

Conference:  IEEE/IFIP NOMS 2020, Budapest, Hungary, 20-24 April 2020

Abstract: In this paper, we examine the problem of the efficient allocation of resources in networking testbeds, which cannot be shared among the experimenters. We highlight the similarities with the housing market where indivisible network resources play the role of houses, while experimenters the role of owners. We adopt the Top-Trading-Cycles (TTC) algorithm for providing Pareto efficient allocations and we compare this approach with the current mechanism of the simple First-Come-First-Served (FCFS) approach used in most networking testbeds. A formulation of the problem is provided where we describe the average utility of the system as a function of the desired testbed resources of the experimenters and the final allocation of the resources to them. In the performance evaluation we observe that TTC outperforms FCFS in all the examined scenarios and achieves almost 95% better average utility in certain cases.

Download paper: [PAPER]

On the Implementation of Matching Theory Based Resource Allocation in Networking Testbeds

Authors: D. Stavropoulos, V. Miliotis, T. Korakis and L. Tassiulas

Conference:  IEEE/IFIP NOMS 2020, Budapest, Hungary, 20-24 April 2020

Abstract: In this paper, we demonstrate the implementation of a matching theory based algorithm that extends the well known Top-Trading-Cycles (TTC) in the resource allocation mechanism of the NITOS networking testbed. We first provide an overview of prominent networking testbeds and the common approach that is used for the allocation of testbed resources to experimenters. We briefly describe our extended TTC algorithm and how it is applied in the enhanced architecture of NITOS resource allocation framework, by illustrating the interactions that take place between the testbed users and the testbed management
services.

Download paper: [PAPER]

More Articles ...

  1. Servicing Inelasticity, Leasing Resources and Pricing in 5G Networks
  2. A Dynamic Pricing and Leasing Module for 5G Networks
  3. EVERUN - Enabling Power Consumption Monitoring in Underwater Networking Platforms
  4. In-situ Power Consumption Meter for Sensor Networks supporting Extreme Dynamic Range