Pierre St. Juste

SocialVPN: Enabling wide-area collaboration with integrated social and overlay networks

Trusted collaborative systems require peers to be able to communicate over private, authenticated end-to-end channels. Network-layer approaches such as Virtual Private Networks (VPNs) exist, but require considerable setup and management which hinder the establishment of ad-hoc collaborative environments: trust needs to be established, cryptographic keys need to be exchanged, and private network tunnels need to be created and maintained among end users. In this paper, we propose a novel system architecture which leverages existing social infrastructures to enable ad-hoc VPNs which are self-configuring, self-managing, yet maintain security amongst trusted and untrusted third parties. The key principles of our approach are: (1) self-configuring virtual network overlays enable seamless bi-directional IP-layer connectivity to socially connected parties; (2) online social networking relationships facilitate the establishment of trust relationships among peers; and (3) both centralized and decentralized databases of social network relationships can be securely integrated into existing public-key cryptography (PKI) implementations to authenticate and encrypt end-to-end traffic flows. The main contribution of this paper is a new peer-to-peer overlay architecture that securely and autonomously creates VPN tunnels connecting social peers, where online identities and social networking relationships may be obtained from centralized infrastructures, or managed in a decentralized fashion by the peers themselves. This paper also reports on the design and performance of a prototype implementation that embodies the SocialVPN architecture. The SocialVPN router builds upon IP-over-P2P (IPOP) virtual networks and a PKI-based tunneling infrastructure, which integrates with both centralized and decentralized social networking systems including Facebook, the Drupal open-source content management system, and emailing systems with PGP support. We demonstrate our prototypes ability to support existing, unmodified TCP/IP applications while transparently dealing with user connectivity behind Network Address Translators (NATs). We also present qualitative and quantitative analyses of functionality and performance based on wide-area network experiments using PlanetLab and Amazon EC2.

On the design of scalable, self-configuring virtual networks

Virtual networks (VNs) provide methods that simplify resource management, deal with connectivity constraints, and support legacy applications in distributed systems, by enabling global addressability of VN-connected machines through either a common layer 2 Ethernet or a NAT-free layer 3 IP network. This paper presents a novel VN design that supports dynamic, seamless addition of new resources with emphasis on scalability in a unified private IP address space. Key features of this system are: (1) Scalable connectivity via a P2P overlay with the ability to bypass overlay routing in LAN communications, (2) support for static and dynamic address allocation in conjunction with virtual nameservers through a distributed data store, and (3) support for transparent migration of IP endpoints across wide-area networks. The approach is validated by a prototype implementation which has been deployed in grid and cloud environments. We present both a quantitative and qualitative discussion of our findings.

Integrating Overlay and Social Networks for Seamless P2P Networking

In this paper we introduce social VPNs, a novel system architecture which leverages existing social networking infrastructures to enable ad-hoc VPNs which are self-configuring, self-managing, yet maintain security against untrusted parties. The key principles in our approach are: (1) self-configuring virtual network overlays enable seamless bi-directional IP-layer connectivity among parties linked by means of social connections; (2) social networking infrastructures greatly facilitate the establishment of trust relationships among parties, and these can be seamlessly integrated with existing public-key cryptography implementations to authenticate and encrypt traffic flows on overlay links end-to-end; and (3) knowledge of social connections can be used to improve the performance of overlay routing. This paper describes the architecture of such Social VPNs and a prototype implementation which integrates the Facebook API, IP-over-P2P virtual networks, and the IPsec security infrastructure in a virtual router. We demonstrate the ability of the prototype to support existing, unmodified TCP/IP applications while transparently dealing with the increasingly common case of users connected to the Internet through network address translators (NATs), and present qualitative and quantitative analysis of its functionality and performance.

Machine Learning-Based Runtime Scheduler for Mobile Offloading Framework

Remote offloading techniques have been proposed to overcome the limited resources of mobile platforms by leveraging external powerful resources such as personal work-stations or cloud servers. Prior studies have primarily focused on core mechanisms for offloading. Yet, adaptive scheduling in such systems is important because offloading effectiveness can be influenced by varying network conditions, workload requirements, and load at the target device. In this paper, we present a study on the feasibility of applying machine learning techniques to address the adaptive scheduling problem in mobile offloading framework. The study considers 19 different machine learning algorithms and four workloads, with a dataset obtained through the deployment of an Android-based remote offloading framework prototype on actual mobile and cloud resources. From this set, a subset of machine learning algorithms, which have relatively high scheduling accuracy, is selected to implement an offline offloading scheduler. Finally, by taking computational cost and the scheduling performance into account, we use Instance-Based Learning to evaluate an online adaptive scheduler for mobile offloading. In our evaluation, we observe that an Instance Learning-based online offloading scheduler selects the best scheduling decision in 87.5% instances, in an experiment setup in which an image processing workload is offloaded while subject to varying network bandwidth conditions and the amount of data transfer.

TinCan: User-Defined P2P Virtual Network Overlays for Ad-hoc Collaboration

Virtual private networking (VPN) has become an increasingly important component of a collaboration environment because it ensures private, authenticated communication among participants, using existing collaboration tools, where users are distributed across multiple institutions and can be mobile. The majority of current VPN solutions are based on a centralized VPN model, where all IP traffic is tunneled through a VPN gateway. Nonetheless, there are several use case scenarios that require a model where end-to-end VPN links are tunneled upon existing Internet infrastructure in a peer-to-peer (P2P) fashion, removing the bottleneck of a centralized VPN gateway. We propose a novel virtual network — TinCan — based on peerto-peer private network tunnels. It reuses existing standards and implementations of services for discovery notification (XMPP), reflection (STUN) and relaying (TURN), facilitating configuration. In this approach, trust relationships maintained by centralized (or federated) services are automatically mapped to TinCan links. In one use scenario, TinCan allows unstructured P2P overlays connecting trusted end-user devices — while only requiring VPN software on user devices and leveraging online social network (OSN) infrastructure already widely deployed. This paper describes the architecture and design of TinCan and presents an experimental evaluation of a prototype supporting Windows, Linux, and Android mobile devices. Results quantify the overhead introduced by the network virtualization layer, and the resource requirements imposed on services needed to bootstrap TinCan links.

SNARF: a social networking-inspired accelerator remoting framework

The diminishing size and battery requirements of mobile devices restrict the scope of computations possible on such devices and motivate approaches that support the selective offloading of computations to remote resources. With a variety of resources available to potentially host offloaded computations -- such as cloud-provisioned resources, and devices within a users personal or social network -- a key challenge lies in architecting a framework that enables applications to seamlessly discover available services, effectively and securely communicate with them, and be presented with API interfaces that hide the complexities associated with managing the interactions with a remote device from applications and present the abstraction of a local device. In this paper, we outline a framework that addresses these challenges by layering APIs and an offload infrastructure upon a virtual networking substrate that supports TCP/IP networking and widely-used resource discovery protocols. An intelligent runtime scheduling layer monitors the execution environment and provides opportunistic remote offloads based on the performance requirements, offload benefits and expendable power. We demonstrate the feasibility of the approach through experiments that evaluate end-to-end application execution times and energy consumption in offloaded mobile devices, as well as the ability to support universal plug-and-play (UPnP) resource discovery in both local- and wide-area environments.

Low Energy Socially Cognizant Routing for Delay Tolerant Mobile Networks

Many message delivery services are based on publish-subscribe systems designed to distribute updates through centralized infrastructures requiring active Internet connections. For mobile devices, individual nodes should have the ability to propagate messages to interested users over ad-hoc wireless connections thereby removing the dependence on Internet and centralized servers. These nodes are sometimes stationary, but are often mobile, creating intermittent networks of nodes that tend to be socially related. In this paper, we propose LESC, a delay-tolerant message delivery protocol, which facilitates efficient message dissemination in a decentralized, ad-hoc fashion and can be implemented using a commodity mobile communication technology such as Bluetooth LE. By leveraging the frequent collocation of socially related peers, nodes strategically become information carriers with the ability to propagate messages to out of range nodes in the future. We design a discrete event simulator that utilizes actual traveling paths derived from Google Maps. The simulator emulates LESC and the epidemic routing protocol to determine if we can achieve reasonable performance. Related works have approached the problem of publish-subscribe systems on mobile devices, but to the best of our knowledge, have not shown the feasibility of a protocol that can directly be implemented over current commodity wireless technologies. We simulate the protocol in Matlab and allow nodes to have multiple publications and subscriptions simultaneously.

Litter: A Lightweight Peer-to-Peer Microblogging Service

Microblogging has become an important part of the social web evolution and is being utilized in many aspects such as advertising, political campaigns, and popular uprisings. Due to its heavy centralization, many have proposed decentralized alternatives based on a variety of models. This paper suggests a fully distributed approach built on top of existing peer-to-peer technologies. We demonstrate that, by exploiting the services of current peer-to-peer middleware along with the properties of the social graph, it is possible to create a simple, yet practical microblogging service that is impervious to many of the shortcomings of their centralized counterparts. The approach has been implemented as a software prototype that is readily available for download in order to test our design in real life environments.

OpenCL-Based Remote Offloading Framework for Trusted Mobile Cloud Computing

OpenCL has emerged as the open standard for parallel programming for heterogeneous platforms enabling a uniform framework to discover, program, and distribute parallel workloads to the diverse set of compute units in the hardware. For that reason, there have been efforts exploring the advantages of parallelism from the OpenCL framework by offloading GPGPU workloads within an HPC cluster environment. In this paper, we present an OpenCL-based remote offloading framework designed for mobile platforms by shifting the motivation and advantages of using the OpenCL framework for the HPC cluster environment into mobile cloud computing where OpenCL workloads can be exported from a mobile node to the cloud. Furthermore, our offloading framework handles service discovery, access control, and data privacy by building the framework on top of a social peer-to-peer virtual private network, Social VPN. We developed a prototype implementation and deployed it into local- and wide-area environments to evaluate the performance improvement and energy implications of the proposed offloading framework. Our results show that, depending on the complexity of the workload and the amount of data transfer, the proposed architecture can achieve more energy efficient performance by offloading than executing locally.We can classify multi-FPGA prototyping platforms in three categories: hardwired off-the-shelf, cabling and custom. Three points are developed in this paper. Firstly, an automatic design flow is proposed to generate a cabling platform and a custom platform for a given design. Then, the optimal width of cables for a cabling multi-FPGA platform is explored. Finally, the performances of these three multi-FPGA platforms are compared. The results show that the cabling platform achieves up to 82% gain in performance, and the custom platform achieves up to 100%, compared to the hardwired off-the-shelf platform. The custom platform achieves up to 20% gain in performance over the cabling platform. Therefore the results show that, apart from some stringent constraints (such as deployment cost or specific frequency needed), the relatively new cabling paradigm with the proposed automatic, inter-FPGA tracks distribution tool, offers an attractive alternative compared to the two other platforms.In this paper we propose the multi-user MIMO (MU-MIMO) with the fixed relay station (RS) transmissions scheme where the base station (BS) and the fixed RS are equipped with multiple antennas and the multiple users are equipped with a single antenna. Since the MU-MIMO can improve the sum capacity significantly by transmitting the data streams to multiple users simultaneously without much link gain while the fixed relay station can improve the link gain but loses the spectral efficiency due to the half-duplex transmission, our proposed scheme can compensate each of the drawbacks and can improve both the sum rate and the link gain. We adopt the practical random unitary beamforming (RUB) precoding method for MU-MIMO and the simple amplify-and-forward protocol for the operation of the relay station. Based on our proposed RUB based MU-MIMO with fixed RS, the sum capacity is upper bounded by the channels between the BS and the RS, not by the channels between the RS and the users. We also show from our extensive simulations that our proposed scheme has lower complexity compartively and provides the performance commensurate with the schemes presented in the literature in terms of the sum-rate for the large user scenario.

OverSoc: Social Profile Based Overlays

Online social networking has quickly become one of the most common Internet activities. As social networks evolve, they encourage users to share more information, requiring the users, in turn, to place more trust into social networks. In centralized systems, this means trusting a third-party commercial entity, like Facebook or MySpace. Peer-to-peer (P2P) systems can enable the creation of online social networks extending trust to friends only. In this paper, we present a novel approach to constructing completely decentralized social networks through P2P overlays, OverSoc. Our approach relies on a common directory overlay, which facilitates friend discovery and bootstraps connectivity to individualized profile overlays. Each user has their own individual profile overlay managed transparently using a public key infrastructure (PKI). We define necessary interfaces for constructing the system and describe some examples of user interactions with the system.