Special issue | Cloud Communications and Networking

Vol. 73, n° 3-4, March-April 2018
Content available on Springerlink 

Guest editors

Harry Perros, NC State University, USA
Ioannis Papapanagiotou, Netflix, USA
Guy Pujolle, UPMC, France


Cloud Communications and Networking

Harry Perros, Ioannis Papapanagiotou, Guy Pujolle

A survey on the communication and network enablers for cloud-based services: state of the art, challenges, and opportunities

George Patrick Xavier, Burak Kantarci

The School of Electrical Engineering and Computer Science, University of Ottawa, Ottawa, Canada

Abstract The wide adoption of the cloud computing concept has resulted in major impacts in both fixed and mobile communication networks leading to cutting-edge research to provide appropriate network architecture and protocols, along with resource management mechanisms. Cloud computing research has been witnessing the interplay between the system and communication aspects in order to offer powerful inter-networking and interoperability between the systems and networks. This paper reviews recent works focusing on architectural design issues, virtualization solutions, and challenges in cloud communications and networking. We mainly discuss the architectural challenges and solutions in today’s leading cloud communication technologies starting with network virtualization, software-defined networking (SDN), network function virtualization (NFV), and SDN-enabled NFV solutions. Furthermore, considering the benefits of cloud computing for mobile communications, we overview the cloud-RAN architecture for radio access networks, along with its support for various existing and future wireless communication technologies including future 5G wireless networks. We study each cloud communication technology by focusing on the existing works from the standpoint of objectives, challenges, and solutions. Furthermore, for all cloud communication concepts, we present a thorough discussion on the open issues and opportunities.

Keywords Cloud communications, Cloud networking, Network virtualization, Software-defined networking, Network function virtualization, Cloud radio access networks, 5G 

AutoSAC: automatic scaling and admission control of forwarding graphs

Victor Millnert1, Enrico Bini2, Johan Eker1,3

(1) Lund University, Lund, Sweden
(2) Università degli Studi di Torino, Torino, Italy
(3) Ericsson Research, Lund, Sweden

Abstract There is a strong industrial drive to use cloud computing technologies and concepts for providing timing sensitive services in the networking domain since it would provide the means to share the physical resources among multiple users and thus increase the elasticity and reduce the costs. In this work, we develop a mathematical model for user-stateless virtual network functions forming a forwarding graph. The model captures uncertainties of the performance of these virtual resources as well as the time-overhead needed to instantiate them. The model is used to derive a service controller for horizontal scaling of the virtual resources as well as an admission controller that guarantees that packets exiting the forwarding graph meet their end-to-end deadline. The Automatic Service and Admission Controller (AutoSAC) developed in this work uses feedback and feedforward making it robust against uncertainties of the underlying infrastructure. Also, it has a fast reaction time to changes in the input.

Keywords Cloud computing, Network function virtualisation, End-to-end deadline, Real-time, Feedback control, Feedforward control  

Dynamic VM allocation in a Saas environment

Brian Bouterse, Harry Perros

Department of Computer Science, North Carolina State University, NC, USA

Abstract Given the costs associated with a cloud infrastructure, dynamic scheduling of virtual machines (VMs) can significantly lower costs while providing an acceptable service level. We develop a series of forecasting models for predicting demand for VMs in a cloud-based software used as a software-as-a-service (SaaS). These models are then used in a periodic-review provision model which determines how many VMs should be provisioned or de-provision at each inspection interval. A simple provisioning heuristic model is also proposed, whereby a fixed reserve capacity of VMs is continuously maintained. We evaluate and compare the performance of these models for different model parameters using historical data from the Virtual Computing Laboratory (VCL) at North Carolina State University.

Keywords Provision of virtual machines (VMs), Time-series forecasting models, Hidden Markov models (HMM), Virtual computing laboratory (VCL), Periodic-review model Capacity planning  

Constrained max-min fair scheduling of variable-length packet-flows to multiple servers

J. Khamse-Ashari1, G. Kesidis2, I. Lambadaris1B. Urgaonkar2, Y. Zhao3

(1) SCE Department, Carleton University, Ottawa, Ontario, Canada
(2) School of EECS, Pennsylvania State University, PA, USA
(3) School of Mathematics and Statistics, Carleton University, Ottawa, Ontario, Canada

Abstract In this paper, we study a multi-server queuing system wherein each user is constrained to get service only from a specified subset of servers. Fair packet scheduling in such a setting poses novel challenges that we address in this paper. Specifically, we observe that max-min fair allocation of the available resource over different servers (notably bandwidth) in the presence of placement constraints results in different levels of fair service-rates. To achieve the max-min fair service rates, we propose a novel packet scheduler which is inspired by the deficit-round robin (DRR) algorithm. The scheduler allocates tokens to flows in a round-by-round manner, where token allocation to flows at the beginning of each round is weighted max-min fair. So, we have called it multi-server max-min fair DRR (MSMF-DRR). The performance of the MSMF-DRR algorithm in terms of achieving fairness is shown through a worst-case performance analysis. In addition to analytical results, numerical experiments are also carried out to illustrate service isolation and the delay guarantee that are provided by the algorithm. Generally, a scheduler for such a constrained multi-server queuing system can be applicable in many modern data-networking applications, especially in cloud computing wherein virtual machines and/or processes vie for different IT resources distributed over heterogenous servers, while different processes may have preferences over servers owing to their quality-of-service requirements and the heterogeneity of servers.

Keywords Packet scheduling, K-server algorithms, Placement constraints, Max-min fairness, Cloud computing, Resource allocation, Convex optimization

SDN-based Wi-Fi Direct clustering for cloud access in campus networks

Thi Mai Trang Nguyen1, Lyes Hamidouche1,2, Fabien Mathieu3Sébastien Monnet4, Syphax Iskounen5

(1) LIP6, Sorbonne Université, UPMC, CNRS, Paris, France
(2) Magency, Paris, France
(3) NOKIA Bell Labs, Nozay, France
(4) LISTIC, Université Savoie Mont Blanc, Anncey le Vieux, France
(5) LASS, CNRS, Toulouse, France

Abstract Mobile cloud is changing the way to enroll teaching activities in a university campus. Lectures and lab sessions can be carried out directly from tablets in a classroom by accessing a server in the cloud. In this paper, we address the problem of high-density cloud access with wireless devices in campus networks. We propose to use Wi-Fi direct clustering to solve the problem of quality of service (QoS) degradation when a high number of wireless devices want to access a content in the cloud at the same time. A centralized software-defined network controller is used in our proposed architecture to capture the network state and organize the Wi-Fi Direct groups. The optimized number of clusters can be calculated in function of the number of devices in the room. By simulations, we show that we can provide a better QoS in terms of download time and application’s throughput by reducing the interference in this dense wireless network environment.

Keywords Wi-Fi Direct, Software-defined networking, Cloud network access

Assignment and collaborative execution of tasks on transient clouds

Mina Guiguis, Qijun Gu, Terry Penner, Lavanya Tammineni, Thomas Langford, Agustin Rivera-Longoria, Alison Johnson, Brandon Van Slyke

Texas State University, San Marcos, TX, USA

Abstract Transient clouds (TC) are temporal clouds that enable nearby mobile devices to form an ad hoc network and advertise their capabilities as cloud services. Through utilizing the collective power of the group, devices are no longer constrained by their local hardware and software capabilities. TC harness the ubiquitous nature of mobile devices along with their ever-increasing sets of capabilities in providing a rich computing platform. In this paper, we present two instantiations of task assignment algorithms that achieve various goals such as balancing the load on devices and minimizing the cost of communication. In the first instantiation, we consider a centralized approach in which a cluster head is responsible for maintaining the list of capabilities and assigning tasks to devices based on their capabilities. We present a modified version of the Hungarian method that allows for balancing the load on devices. In the second instantiation, we consider a distributed approach in which devices advertise and find capabilities through an overlay network. The overlay is designed to capitalize on locality and thus seeks to minimize the cost in finding devices with certain capabilities. We evaluate the performance of our TC through extensive simulation experiments complemented by a realistic implementation on a set of devices.

Keywords Mobile cloud computing, Task assignment, Peer-to-peer networking

Mitigating incast-TCP congestion in data centers with SDN

Ahmed M. Abdelmoniem, Brahim Bensaou, Amuda James Abu

Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Kowloon, Hong Kong

Abstract In data center networks (DCNs), the presence of long-lived TCP flows tends to bloat the switch buffers. As a consequence, short-lived TCP-incast traffic suffers repeated losses that often lead to loss recovery via timeout. Because the minimum retransmission timeout (minRTO) in most TCP implementations is fixed to around 200 ms, interactive applications that often generate short-lived incast traffic tend to suffer unnecessarily long delays waiting for the timeout to elapse. The best and most direct solution to such problem would be to customize the minRTO to match DCNs delays; however, this is not always possible; in particular in public data centers where multiple tenants, with various versions of TCP, co-exist. In this paper, we propose to achieve the same result by using techniques and technologies that are already available in most commodity switches and data centers and that do not interfere with the tenant’s virtual machines or TCP protocol. In this approach, we rely on the programmable nature of SDN switches and design a SDN-based incast congestion control (SICC) framework, that uses a SDN network application in the controller and a shim-layer in the host hypervisor, to mitigate incast congestion. We demonstrate the performance gains of the proposed scheme via real deployment in a small-scale testbed as well as ns2 simulation experiments in networks of various sizes and settings.

Keywords Congestion control, Data center networks, Incast, Software-defined networking, TCP

Comiqual: collaborative measurement of internet quality

Marc Ibrahim, Maroun Chamoun, Rima Kilany, Melhem El Helou, Nicolas Rouhana

Ecole Supérieure d’Ingénieurs de Beyrouth, Saint Joseph University of Beirut, Beirut, Lebanon

Abstract With the continuous growth of both fixed and mobile Internet usage, measuring the Internet QoS (quality of service) becomes of vital interest for all involved Internet stakeholders, mainly consumers, operators, and regulators. In this paper, we describe in detail, Comiqual (collaborative measurement of Internet quality), a crowd-sourced large-scale Internet measurement platform that coordinates and collects measurements from measurement agents (MAs) installed on fixed and mobile end user devices. Although the initial and main target of Comiqual is Lebanon, the platform is generically designed to measure the Internet access quality from the user’s perspective anywhere on the globe. The MAs that execute mainly active measurements are jointly controlled by users and by a measurement center (MC); the latter sends measurement instructions to MAs and collects the measurement results. The communication protocol between MC and MAs uses JSON messages that are exchanged via HTTP through REST calls and secured by HTTPS. Measurement results could be openly accessed in a raw format or viewed as an aggregation via a Google map. Moreover, an online statistical tool allows user-defined statistics computation and visualization. All these features combined with the flexibility of the platform management are the main drivers that will allow Comiqual to reach its ultimate goal, which is to create a collaborative, neutral, and transparent observatory of the Internet.

Keywords Internet, Crowd-sourcing, Large-scale measurement, Network performance, Open data