Vinicio Vercellone

Design considerations for a network of information

The existing Internet ecosystem is a result of decades of evolution. It has managed to scale well beyond the original aspirations. Evolution, though, highlighted a certain degree of inadequacies that is well documented. In this position paper we present the design considerations for a re-architected global networking architecture which delivers dissemination and non-dissemination objects only to consenting recipients, reducing unwanted traffic, linking information producers with consumers independently of the hosts involved, and connects the digital with the physical world. We consider issues ranging from the proposed object identifier/locator split to security and trust as we transition towards a Network of Information and relate our work with the emerging paradigm of publish/subscribe architectures. We introduce the fundamental components of a Network of Information, i.e., name resolution, routing, storage, and search, and close this paper with a discussion about future work.

MDHT: a hierarchical name resolution service for information-centric networks

Information-centric network architectures are an increasingly important approach for future Internet architectures. Several approaches are based on a non-hierarchical identifier (ID) namespace that requires some kind of global Name Resolution Service (NRS) to translate the object IDs into network addresses. Building a world-wide NRS for such a namespace with 1015 expected IDs is challenging because of requirements such as low latency, efficient network utilization, and anycast routing. In this paper, we present an NRS called Multi-level Distributed Hash Table (MDHT). It provides name-based anycast routing, can support constant hop resolution, and fulfills the afore mentioned requirements. A scalability assessment shows that our system can scale to the Internet level, managing 1015 objects with todays storage technology and 1/10th of todays DNS nodes. The evaluation indicates that a non-hierarchical namespace can be adopted on a global scale, opening up several design alternatives for information-centric network architectures.

Hierarchical DHT-based name resolution for information-centric networks

Information-centric network (ICN) architectures are an increasingly important approach for the future Internet. Several ICN approaches are based on a flat object ID namespace and require some kind of global name resolution service to translate object IDs into network addresses. Building a world-wide NRS for a flat namespace with 10^1^6 expected IDs is challenging because of requirements such as scalability, low latency, efficient network utilization, and anycast routing that selects the most suitable copies. In this paper, we present a general hierarchical NRS framework for flat ID namespaces. The framework meets those requirements by the following properties: The registration and request forwarding matches the underlying network topology, exploits request locality, supports domain-specific copies of binding entries, can offer constant hop resolution (depending on the chosen underlying forwarding scheme), and provides scoping of publications. Our general NRS framework is flexible and supports different instantiations. These instantiations offer an important trade-off between resolution-domain (i.e. subsystem) autonomy (simplifying deployment) and reduced latency, maintenance overhead, and memory requirements. To evaluate this trade-off and explore the design space, we have designed two specific instantiations of our general NRS framework: MDHT and HSkip. We have performed a theoretical analysis and a simulation-based evaluation of both systems. In addition, we have published an implementation of the MDHT system as open source. Results indicate that an average request latency of (well) below 100ms is achievable in both systems for a global system with 12 million NRS nodes while meeting our other specific requirements. These results imply that a flat namespace can be adopted on a global scale, opening up several design alternatives for information-centric network architectures.

Providing Data Dissemination Services in the Future Internet

In this paper authors examine benefits and motivations for the Information-Centric Networking (ICN) concept [1][2] under the point of view of an Internet Service Provider (ISP). A proposal for a Global Information Network (GIN) is advanced, which should follow a standardization path. Authors believe that ICN will enable ISPs to provide native support to typical data dissemination services (e.g. file sharing, distributed storage, caching, instant messaging, VoIP, etc.) currently provided by overlay networks and P2P applications (e.g. eMule/eDonkey [3], BitTorrent [4], Kazaa [5], Skype [6], Akamai [7], etc.). This will be possible by integrating most of the basic features of the current P2P overlay networks into the ISPs network infrastructure. As a result, new advanced communication services will be offered at a new networking layer, sitting on top of a transport infrastructure, based either on IP or whatever new transport technology. The potential of this change in the Internet architecture could be dramatic, since ICN impacts both the communication model and the network infrastructure. Furthermore, it is foreseeable that the business relationships among most of the actors (customers, network providers, ISPs, content providers, search engines) will be also deeply changed.

Introducing network-aware scheduling capabilities in OpenStack

This paper motivates and describes the introduction of network-aware scheduling capabilities in OpenStack, the open-source reference framework for creating public and private clouds. This feature represents the key for properly supporting the Network Function Virtualization paradigm, particularly when the physical infrastructure features servers distributed across a geographical region. This paper also describes the modifications required to the compute and network components, Nova and Neutron, and the integration of a network controller into the cloud infrastructure, which is in charge of feeding the network-aware scheduler with the actual network topology.

Toward dynamic virtualized network services in telecom operator networks

NFV and SDN are nowadays seen as a solid opportunity by telecom operators to reduce costs while at the same time providing new and better services. Recently, the Unify project proposed a multi-layered architecture that, leveraging different levels of abstraction, can orchestrate and deploy generic network services on the physical infrastructure of the telecom operator. In this paper, we exploit such an architecture to deliver end-to-end generic services in presence of multiple concurring players (e.g. network operator, end-users), leveraging a new simple data model. Particularly, we propose a description-based approach allowing the deployment agile, implementation-independent and high-level network services over a distributed set of resources. The resulting data model can abstract generic services, including both middlebox-based (e.g., firewalls, NATs, etc.) and traditional LAN-based ones (e.g., a BitTorrent client). Finally, two distinct prototypes, originated by different design principles, are implemented in order to validate our proposal with the aim of demonstrating the adaptability of our approach to different contexts.

Scalability of ONOS reactive forwarding applications in ISP networks

Software Defined Networking (SDN) is a powerful approach that enhances network control and management, and provides a flexible way to develop network applications. However, scalability of SDN networks is an important concern for many network operators. The main peculiarities of SDN when applied to an Internet Service Provider (ISP) network are the large geographical extension and the need of in-band transmission of control traffic. Therefore, the control traffic exchanged between the SDN controller and the network nodes must be carefully evaluated for the network design and dimensioning.We consider an ISP network controlled by the recently ONOS (Open Network Operating System) controller developed by ON.Lab. We devise a quantitative model to compute the exact number of exchanged OpenFlow messages and the corresponding bandwidth needed to install a traffic flow when running the default ONOS layer-2 forwarding applications. We compute also the exact number of flow rules installed in each switch. We show the general applicability of our models for a Point Of Presence (POP) network and for a large set of real nation-wide and world-wide ISP networks.Our quantitative models can be used for a safe network planning also when the network applications are not fully reactive.

Evaluating the SDN control traffic in large ISP networks

Scalability of Software Defined Networking (SDN) approach is one of the key issues that many network operators are willing to address and understand. Indeed, the promised programmability and flexibility of a SDN network is paid with a non-negligible control traffic exchanged between the network nodes and the SDN controllers. We consider a network controlled by a real OpenFlow-enabled controller, i.e. OpenDaylight. We evaluate analytically the number of OpenFlow messages for installing a new traffic flow, assuming the default reactive forwarding application available in OpenDaylight. We apply these results to the specific case of a large ISP network, comprising a backbone interconnecting many POPs. By evaluating exactly the amount of generated control traffic, we are able to assess the scalability of the reactive forwarding application in a practical relevant scenario for ISPs.

Per-user NFV services with mobility support

This paper presents an architecture to provide endto- end per-user services with support to client mobility, designed according to the SDN and NFV paradigms. Our service platform dynamically configures and launches service requests when the client connects to the network, which are used by a multidomain orchestration system to arrange the required network configuration and computational resources. Service configuration is dynamically updated when a movement of the client is detected, that is, when a client device changes its access point to the network. A prototype implementing the idea has been developed and validated over JOLNET, a real, geographical, OpenFlowbased experimental network connecting several sites in Italy and operated by Telecom Italia.

Design and implementation of locality-aware P2P system

One of the most relevant problem for an Internet Service Provider is the large bandwidth usage on international links, mainly due to peer-to-peer applications adopted for file-sharing. The Collaborative Locality-aware Overlay SERvice (CLOSER) technology has been recently proposed to solve this issue by properly modifying the behavior of peer-to-peer application. The technology is also covered in two recent patent applications. This paper presents possible design guidelines to actually implement CLOSER in a DHT-based peer-to-peer system and describe a real implementation based on the popular aMule application