Dávid Jocha

Scalable fault management for OpenFlow

In the OpenFlow based split architecture, data-plane forwarding is separated from control and management functions. Forwarding elements make only simple forwarding decisions based on flow table entries populated by the controller. While OpenFlow does not specify how topology monitoring is performed, the centralized controller can use Link-Layer Discovery Protocol (LLDP) messages to discover link and node failures and trigger restoration actions. This monitoring and recovery model has serious scalability limitations because the controller has to be involved in the processing of all of the LLDP monitoring messages. For fast recovery, monitoring messages must be sent with millisecond interval over each link in the network. This poses a significant load on the controller. In this paper we propose to implement a monitoring function on OpenFlow switches, which can emit monitoring messages without posing a processing load on the controller. We describe how the OpenFlow 1.1 protocol should be extended to support the monitoring function. Our experimental results show that data plane fault recovery can be achieved in a scalable way within 50 milliseconds using this function.

Towards Unified Programmability of Cloud and Carrier Infrastructure

The rise of cloud services poses considerable challenges on the control of both cloud and carrier network infrastructures. While traditional telecom network services rely on rather static processes (often involving manual steps), the wide adoption of mobile devices including tablets, smartphones and wearables introduce previously unseen dynamics in the creation, scaling and withdrawal of new services. These phenomena require optimal flexibility in the characterization of services, as well as on the control and orchestration of both carrier and cloud infrastructure. This paper proposes a unified programmability framework addressing: the unification of network and cloud resources, the integrated control and management of cloud and network, the description for programming networked/cloud services, and the provisioning processes of these services. In addition proofs-of-concept are provided based on existing open source control software components.

Elastic network functions: opportunities and challenges

Network function virtualization (NFV) and software defined networking (SDN) are key technology enablers for cost reductions and new business models in networking. The possibility to automatically and dynamically scale network services at run time is one of the main claims of NFV. Elastic NFV could be similar to what elastic cloud services provide for compute, with pay-per-use cost models for customers. However, control of resources for elastic services is far from trivial. We show how current NFV and SDN architectures could support elastic resource services for network functions (NFs). We reveal that the current NFV architecture does not allow recursive resource orchestration, therefore preventing resource scaling requests from being handled by a resource orchestrator overseeing the entire domain where an NF is executed. We introduce a logical centralization of joint compute and network resource orchestration as a UNIFY framework, which enables direct control of elastic resources for the NFs. We show opportunities and challenges associated with such an architecture.

Multi-Domain Service Orchestration Over Networks and Clouds: A Unified Approach

End-to-end service delivery often includes transparently inserted Network Functions (NFs) in the path. Flexible service chaining will require dynamic instantiation of both NFs and traffic forwarding overlays. Virtualization techniques in compute and networking, like cloud and Software Defined Networking (SDN), promise such flexibility for service providers. However, patching together existing cloud and network control mechanisms necessarily puts one over the above, e.g., OpenDaylight under an OpenStack controller. We designed and implemented a joint cloud and network resource virtualization and programming API. In this demonstration, we show that our abstraction is capable for flexible service chaining control over any technology domains.

UNIFYing Cloud and Carrier Network Resources: An Architectural View

Cloud networks provide various services on top of virtualized compute and storage resources. The flexible operation and optimal usage of the underlying infrastructure are realized by resource orchestration methods and virtualization techniques developed during the recent years. In contrast, service deployment and service provisioning in carrier networks have several limitations in terms of flexibility, scalability or optimal resource usage as the built-in mechanisms are strongly coupled to the physical topology and special purpose hardware elements. Network Function Virtualization (NFV) opens the door between cloud and carrier networks by providing software-based telecommunication services which can run in virtualized environment on general purpose hardwares. Our main goal is to unify software and network resources in a common framework. In this paper, we propose a novel architecture supporting automated, dynamic service creation based on a fine-granular service chaining model, SDN and cloud virtualization techniques. First, we introduce the architecture with the main components. Second, the most important benefits are highlighted and compared to other state-of-the-art approaches. Finally, preliminary experiences with our proof-of-concept prototypes are presented.

Scalable Architecture for Service Function Chain Orchestration

Network Function Virtualization (NFV) enables to implement network functions in software, high-speed packet processing functions which traditionally are dominated by hardware implementations. Virtualized Network Functions (NFs) may be deployed on generic-purpose servers, e.g., in datacenters. The latter enables flexibility and scalability which previously were only possible for web services deployed on cloud platforms. The merit of NFV is challenged by control challenges related to the selection of NF implementations, discovery and reservation of sufficient network and server resources, and interconnecting both in a way which ful fills SLAs related to reliability and scalability. This paper details the role of a scalable orchestrator in charge of finding and reserving adequate resources. The latter will steer network and cloud control and management platforms to actually reserve and deploy requested services. We highlight the role of involved interfaces, propose elements of algorithmic components, and will identify major blocks in orchestration time in a proof of concept prototype which accounts for most functional parts in the considered architecture. Based on these evaluations, we propose several architectural enhancements in order to implement a highly scalable network orchestrator for carrier and cloud networks.

Multi-layered Service Orchestration in a Multi-domain Network Environment

In this demo, we show a novel method to multi-layer service orchestration in a multi-domain network. This method is a basic implementation of the three layered concept with multi-layer orchestration designed by the UNIFY project. A global orchestrator is capable of instantiating service elements, i.e., virtual network functions (VNFs), in separate domains. Dedicated local orchestrators indifferent infrastructure domains are responsible for setting up new VNF instances and configuring the underlying network. Our implementation is based on the ESCAPE prototyping framework and an OpenStack (OS) data center with the OpenDaylight (ODL) controller.

Scalability aspects of centralized control of MPLS access/aggregation network

We apply the concept of centralized control to the MPLS access/aggregation network segment of service provider networks. We describe a numerical model to capture the scale and scope of the access/aggregation network and explain the corresponding scalability issues. We compare this numerical model to the experimental results obtained on a test-bed with OpenFlow 1.1 compatible central controller and forwarding elements. We show that the central controller implementation we have mostly matches the performance requirements of the access/aggregation networks for different scenarios.

Efficiency of frequency duplexing used for interference elimination

Careful planning of wireless networks is essential to provide reliable communication, especially that microwave links are frequently chosen to interconnect access points within networks. The optimization of such networks belongs to the frequency assignment problems and focuses on setting polarization of the links or allocating the channels in a frequency duplex to reduce or eliminate the interference that might occur between links. In this paper we are aiming to reveal the unexpected difference between these two well-studied problems if the sharpness of the applied beam widths is relaxed. We provide graph theory based analysis and reveal the main differences between frequency duplexing and polarization setting from mathematical point of view. We give numerical bounds for efficiency and show traps of algorithmic solutions working with interference graphs.

The orchestration in 5G exchange — A multi-provider NFV framework for 5G services

The goal of the 5G Exchange project is to enable cross-domain orchestration of services over multiple administrations. The system we build allows the end-to-end integration of heterogeneous resource and service elements of a multi-vendor technology environment from multiple operators by sharing their network and compute infrastructures via NFV orchestration. We will run an industry control 5G use-case, where one of the VNFs is offered by a 3rd party solution provider as a VNFaaS. We will show i) full automation for end-to-end network service orchestration over multi-provider NFV and VNFaaS offerings with latency and high availability constraints; ii) actor-role models and business interactions and iii) how — with a feedback loop to lifecycle management — the system can adapt to changes.