August 23, 2019
Salle Shannon (a.105) - Maxwell building
Software-Defined Systems for Network-Aware Service Composition and Workflow Placement by Pradeeban Kathiravelu
pour l’obtention du grade de Docteur en sciences de l’ingénieur et technologie
Composing complex workflows efficiently from diverse services on the Internet requires communication and coordination across heterogeneous execution environments, ranging from data centers and clouds to the edge managed by different infrastructure providers. Through complete virtualization of network and its services, network softwarization provides efficient management of network architecture. This dissertation exploits the flexibility and management benefits of the network softwarization to solve the problems of service composition and workflow placement at Internet scale. We present two main contributions: first, a set of extensions to network softwarization to simplify and enhance application development and deployment, and second, a scalable architecture to compose service chains in wide area networks. Finally, we evaluate these contributions in the context of big data applications. We thus intend to mitigate the challenges concerning resource management and interoperability of heterogeneous infrastructures, to efficiently compose and schedule various service workflows at Internet scale, while sharing the network and the computing resources among several users.
Network Softwarization revolutionizes the network landscape in various stages, from building, incrementally deploying, and maintaining the environment. Software-Defined Networking (SDN) and Network Functions Virtualization (NFV) are two core tenets of network softwarization. SDN offers a logically centralized control plane by abstracting away the control of the network devices in the data plane. NFV virtualizes dedicated hardware middleboxes and deploys them on top of servers and data centers as network functions. Despite its growing application, network softwarization has not been fully exploited for effectively composing service workflows of multiple users sharing third-party network infrastructures and services. To this end, we propose our contributions to extend network softwarization for network-aware service composition and workflow placement in heterogeneous infrastructures.
First, we separate network from infrastructure by exploiting network softwarization to move out of data centers toward the edge seamlessly, and from simulations to actual deployments, with little or no additional development effort. We extend SDN in cloud and data center environments to unify various phases of development, by uniformly managing the executions of the network applications from an extended SDN controller, regardless of the execution environment and phase. We thus deploy the workloads seamlessly across the phases, from simulations and emulations to physical deployment environments. We further extend this work to support multiple Service Level Agreements (SLAs) across diverse network flows in data centers, by selectively enforcing redundancy on the network flows. Thus, we aim for Quality of Service (QoS) and efficient resource provisioning, while adhering to user policies. Finally, we design a cloud-assisted overlay network, as a latency-aware virtual connectivity provider. Consequently, we propose cost-efficient data transfers and workflow executions at Internet scale.
Second, we propose a scalable architecture to compose service chains in wide area networks efficiently. We exploit SDN and Message-Oriented Middleware (MOM) for a logically centralized composition and execution of service workflows. We thus propose a Software-Defined Service Composition (SDSC) framework for web service compositions, Network Service Chains (NSCs), and a network-aware execution of data services. We further present Software-Defined Systems (SDS) consisting of virtual network allocation strategies for multi-tenant service executions in large-scale networks comprised of multiple domains.
Finally, we investigate how our proposed SDS can operate efficiently for real-world application scenarios of heterogeneous infrastructures. While traditionally web services are built following standards and best practices such as Web Services Description Language (WSDL), network services and data services offered by different service providers often fall short in providing common Application Programming Interfaces (APIs), thus resulting in vendor lock-in. We look into facilitating interoperability across service implementations and deployments, to enable seamless workflow executions and service migrations. We propose big data applications and smart environments such as Cyber-Physical Systems (CPS) and the Internet of Things (IoT) as our two application scenarios. We thus build CPS and big data applications as composable service chains, offering them an interoperable execution.
Our research contributions highlight that network softwarization can be used to build and deploy network applications with minimal repetitive effort, from initial design and development stages to production. Evaluations on the proposed SDS demonstrate performance and economic benefits to service composition and workflow placement at various scales, from data centers to the Internet. By managing and leveraging redundancy in the network flows and network paths, our SDS prototypes ensure that SLAs are met in the critical network flows of multi-tenant systems. Furthermore, our SDS framework reduces Internet latency by up to 30%, yet in an economic approach. Finally, we elaborate the broader applicability of our proposed SDS by extending it to CPS and big data applications.
Jury members :
- Prof. Luis Veiga (IST/ULisboa, Portugal), supervisor
- Prof. Peter Van Roy (UCLouvain), supervisor
- Prof. Charles Pecheur (UCLouvain), chairperson
- Prof. Etienne Rivière (UCLouvain), secretary
- Prof. Fernando Mira da Silva (IST/ULisboa, Portugal)
- Prof. Javid Taheri (Karlstad University, Sweden)
- Prof. Elhadj Benkhelifa (Staffordshire University, UK)