A QoS-Enhanced Model for Inter-Site Backup Operations in Cloud SDN

Introduction

Data centers backup is a very important operation in order to secure data from being lost. According to the environment and the used applications, there are multiple backup strategies. One significant strategy depends on backing up data between different data centers (located in different geographical locations), which is called inter-site data centers backup. Generally, backup operations use network connections just during the data transfer. However, inter-site connections are not utilized in the rest of the time. In other words, using a reserved bandwidth for inter-site backups leads to waste many resources when the connection is idle.

Cloud environment helps in optimizing the usage of computing and networking resources (Rittinghouse & Ransome, 2016). When the inter-site network connections are managed by a cloud provider, the network can be configured to support multiple users according to their requirements. However, performing backup operations for customers requires changes in the network configuration according to backups’ schedules and the selected paths. This reconfiguration process helps in performing backups in optimized conditions, which is a hard administration task in traditional networks.

Software Defined Networking (SDN) is a new programmable networking paradigm that depends on decoupling the control and the data planes in a network, which gives the network administrator more capabilities for managing and controlling the network (Neto, 2011). This aim was hard to be achieved in the past because of the low processing capabilities and the low network connectivity speed, which makes it hard to centralize the controller. Nowadays this goal has already been achieved by SDN that eases the network management process. The flexibility and the dynamicity of SDN management is significant to face networking performance concerns in data centers, especially with different requirements for many customers. Moreover, as data centers and communications are responsible for a significant amount of the consumed energy by ICT (Information and Communications Technology), SDN helps in reducing the power consumption rate by applying an efficient strategy, whereas the legacy network elements consume more electricity when they stay always in “on” mode (AlSous & Marx Gómez, 2016) (Gallagher, 2017). The nature of the SDN indicates the need for a systems approach that is able to use the SDN’s advantages in a flexible and easy way.

This paper presents an extensive literature review about the related topics and shows the challenges that face the offered network services by cloud providers. There is a real need to enhance the backup services in cloud SDN environment: performance, dynamicity and Quality of Service (QoS). The proposed approach aims to use SDN in cloud networks between data centers, in order to provide an inter-site backup service for customers according to their requirements and the current network infrastructure, which helps in the disaster recovery process. This paper presents a model for performing backup operations more efficiently in a Cloud-SDN environment. After measuring the current QoS-state of the network devices/links and their effects, the model generates a function using Fuzzy-Logic-Rules that helps in selecting appropriate paths for performing inter-site backup operations for different customers. Moreover, the proposed model has been applied as a module in an SDN controller, which is easy to be integrated with other SDN networks. The module is responsible for choosing the cost-efficient path to enhance the inter-site connectivity during backups and to support other cloud services in other times, which in turn helps in enhancing the network utilization.

The research includes a literature review about the main concepts. Then, it presents a summary about the related conducted researches. After that, the proposed model with the development environment are explained. Finally, the test results of applying the model were introduced with an extensive analysis that shows effects of each QoS criteria on the whole backup performance.