Gandi SAS

France

Gandi SAS

France
SEARCH FILTERS
Time filter
Source Type

Del Piccolo V.,GANDI SAS | Del Piccolo V.,Paris-Sorbonne University | Amamou A.,GANDI SAS | Vidalenc B.,Thales Alenia | And 2 more authors.
2016 IEEE Global Communications Conference, GLOBECOM 2016 - Proceedings | Year: 2016

Cloud computing is expected to keep growing and therefore more and more data centers are being built. In parallel of this increasing amount of data centers, there are also more cloud providers. On one side, this heterogeneity in the cloud market allows a user to choose and change its cloud provider. On the other side, thanks to this growing number of cloud providers, it is now possible to have small cloud providers forming a cloud federation. This way, they can mutualize their resources in order to accommodate their clients and compete with bigger cloud providers. However, solutions are needed to interconnect data centers. Those solutions must allow a cloud provider to interconnect its data centers to other cloud providers' one. This interconnection must be achieved even if the data centers do not use the same technology. This is the problem we focus on in this paper and we propose a solution to interconnect a TRILL data center with an OpenFlow one. We design and implement this solution in order to analyze its impact on both throughput and latency. Concerning the latency, we found that our solution increases it by 3% compared to the OpenFlow solution. Throughput-wise, we have an 8% lower throughput than TRILL. © 2016 IEEE.


Cerovic D.,GANDI SAS | Del Piccolo V.,GANDI SAS | Amamou A.,GANDI SAS | Haddadou K.,GANDI SAS
2016 3rd Smart Cloud Networks and Systems, SCNS 2016 | Year: 2017

Packet processing speeds in software today can hardly support the speed available on network interfaces. Consequently, many solutions that try to accelerate the processing by using programmable hardware have appeared recently. TRILL protocol is implemented by devices called RBridges which introduce Layer 3 routing features into link layer and serve for communication within a data center. However connection throughput might be affected due to the fact that on each hop, processing of the TRILL header needs to be done in order to decide what is the next hop MAC address. Therefore, our idea was to offload the TRILL data plane on the programmable card in order to improve the network performance, mainly to increase the throughput and to decrease the delay. For this purpose we use Massively Parallel Processor Array (MPPA) which integrates 256 processing engine (PE) cores distributed across 16 compute clusters and use a programming model based on POSIX (Portable Operating System Interface) threads. © 2016 IEEE.


News Article | October 28, 2016
Site: www.marketwired.com

MENLO PARK, CA--(Marketwired - October 24, 2016) - ONOS Project, the rapidly growing open source community advancing the software-defined networking (SDN) OS for service providers with high availability, scale and performance and the right abstractions to create apps and services, today announced BII, Canonical, NCTU and UPMC have joined as collaborators. To build on the continued project momentum, more than 215 ONOS® developers focused on coding and testing will converge in Paris from November 2-4 for the project's first large-scale developer summit, ONOS Build 2016. "ONOS Build 2016 will offer the unique opportunity for ONOS users and contributors across the globe to meet, align, plan and hack together in-person," said Bill Snow, VP of engineering for ON.Lab. "Open to members and non-members alike, attendees will have the chance to talk directly with the ONOS core architects and Technical Steering Team, collaborate on ideas that will directly impact the future of ONOS, and promote their work among the community." The ONOS Build 2016 agenda includes a mix of sharing the ONOS roadmap, social activities for the community to bond, and hacking activities to help deliver specific features in the pipeline. The event will feature keynotes and industry panelists from ONOS community partners and collaborators Ciena, Create-Net, DT, GARR, Huawei, NOKIA Bell Labs, NTT Communications, ON.Lab, Radisys and UPMC. Topics for discussion will include basics of the ONOS architecture, northbound and southbound protocols, deployments, application development, performance and testing, community planning and the Ambassador program. Click here to learn more about the confirmed speakers. A separate Community Showcase track will be devoted to members and ONOS Brigade teams presenting updates on interesting new features developed by and for the ONOS community through open source collaboration. For example, the Dynamic Configuration Brigade has focused on building a vendor-agnostic driver into ONOS that allows automatic discovery and activation of NETCONF-enabled devices and services into the network as long as they expose a Yang Model. Click here to read more about the ONOS Brigades, and here to view the complete ONOS Build 2016 agenda. To obtain an ONOS Build ticket, complete the online registration form. ONOS Build 2016 is free for students and ONOS contributors. General admission costs $400 USD. ONOS is grateful for its generous sponsors who are helping organize the event. These include Platinum sponsors Ciena, Fujitsu, and Huawei, Gold sponsors Kisti and Radisys, and Silver sponsors ADARA Networks and Gandi. If interested in sponsoring ONOS Build 2016 or future Build events, please contact onos-build@onlab.us. More about the new ONOS collaborators: BII: BII Group Holdings Ltd. centered on the core of technical research and development, seated on the basis of testing service, and navigated by internationalization and marketization, is a global open platform for internet infrastructure technologies with a focus on IPv6, Domain Name System (DNS), SDN and Internet of Things (IoT). Canonical: Canonical is the company behind Ubuntu, the leading OS for container, cloud, scale-out and hyperscale computing. 65% of large-scale OpenStack deployments are on Ubuntu, using both KVM and the pure-container LXD hypervisor for the world's fastest private clouds. Canonical provides enterprise support and services for commercial users of Ubuntu. NCTU: National Chiao Tung University (NCTU) is a prestigious university known for its special strengths in the areas of computer science and electronics engineering. Its highly-reputable computer science department, which is the largest one in Taiwan, will participate in the ONOS Project with a focus on SDN-IP. UPMC: The University Pierre and Marie Curie LIP6 computer science research laboratory is dedicated to application modeling and testing, as well as implementation and validation through academic and industry partnerships. Currently, LIP6 is strategizing data consistency for distributed network control-plane systems, and enhancing SDN southbound interfaces. Redefining network economics, ONOS provides the only SDN control plane that can support both disruptive and incremental SDN for service providers and enterprises seeking to virtualize and optimize their networks to keep agile pace with the explosion of mobile devices, video and big data applications. The rapidly growing and diverse ONOS community comprises a core engineering team at ON.Lab, along with developers from service providers, vendors and Research and Educational Networks spanning across industries. Whether an individual or an organization, as an open source project all are encouraged to get involved with the growing ONOS community and help contribute to the project today. ONOS® is the open source SDN networking operating system for Service Provider networks architected for high performance, scale and availability. The ONOS ecosystem comprises ON.Lab, organizations that are funding and contributing to the ONOS initiative, and individual contributors. These organizations include AT&T, China Unicom, Comcast, Google, NTT Communications Corp., SK Telecom Co. Ltd., Verizon, Ciena Corporation, Cisco Systems, Inc., Ericsson, Fujitsu Ltd., Huawei Technologies Co. Ltd., Intel Corporation, NEC Corporation, Nokia, Radisys and Samsung. See the full list of members, including ONOS' collaborators, and learn how you can get involved with ONOS at onosproject.org. ONOS is an independently funded software project hosted by The Linux Foundation, the nonprofit advancing professional open source management for mass collaboration to fuel innovation across industries and ecosystems.


Bourguiba M.,University Paris - Sud | Agha K.A.,University Paris - Sud | Haddadou K.,Gandi SaS
2012 3rd International Conference on the Network of the Future, NOF 2012 | Year: 2012

Mobile cloud computing uses cloud techniques for data storage and processing on behalf of mobile customers. It mitigates the mobile devices limitations in terms of battery and processing capacities. Virtualization is levereged in mobile clouds to enhance resource availability. However, it introduces an overhead that significantly harms the applications responsiveness. In this paper we propose an architecture for a mobile cloud where the mobile devices as well as the access network equipments are integrated in the cloud infrastructure through virtualization. The proposed architecture is enhanced with a novel scheme for network I/O virtualization. Through experimental performance evaluation, we show that our proposal considerably improves mobile devices networking performance in terms of delay and bandwidth when accessing virtual resources in the cloud. © 2012 IEEE.


Bourguiba M.,University Paris - Sud | Haddadou K.,Gandi SaS | Korbi I.E.,National School in Computer Science | Pujolle G.,University Pierre and Marie Curie
IEEE Transactions on Parallel and Distributed Systems | Year: 2014

Virtualization is a key technology to enable cloud computing. It enhances resource availability and offers high flexibility and cost effectiveness. However, the driver domain-based model for network I/O virtualization exhibit poor networking performance. In this paper, we evaluate the networking performance of VMs based on Xen. The I/O channel transferring packets between the driver domain and the VMs is shown to be the bottleneck. To overcome this limitation, we proposed a packet aggregation-based mechanism to transfer packets from the driver domain to the VM. Experimental performance evaluation showed that our proposal allows the VMs throughput to scale up at line rates. Furthermore, we studied the impact of the aggregation on the packets delay. For this purpose, we proposed a system modeling based on queuing theory. The proposed model allowed us to dynamically tune the aggregation mechanism to achieve the best tradeoff between the packets delay and throughput. The proposed I/O virtualization model henceforth satisfies the infrastructure providers to offer cloud computing services. © 2014 IEEE.


It's official — Popcorn Time's torrent streaming service has permanently gone offline, but the Popcorn Time team is not done yet. After learning from the first site, the developers are planning to roll out a legal alternative called Project Butter. Just a few days ago, the hubbub started because of an internal dispute in the Popcorn Time team. With a potential lawsuit looming, several of the core developers for the torrent-streaming site left the team. When the person in charge of the domain name left, the .io domain ceased to work. At one point, the site was even brought back up, but only for a short period. In an attempt to push through with the original project, the Popcorn Time team intended to transfer the domain to a new owner. But Gandi.net, the domain's service provider, withdrew the changes. "In the last few days someone has been tampering with our infrastructure, mainly our DNS service and we can't convince our provider https://gandi.net that we are us and want to stay online," the team said before the Popcorn Time website went offline due to DNS attacks. The team, however, is not yet finished and intends to make a comeback with Popcorn Time's legal counterpart: Project Butter. "This new project, hosted on GitHub, is basically your beloved Popcorn Time stripped down of the parts that made people wary," the new Project Butter team writes. According to the developers, when this new project goes live, it will "never stream any copyrighted material." Project Butter is essentially Popcorn Time minus the direct links to pirated TV shows and movies. Although Popcorn Time's younger brother will be somewhat restricted, it is hoped that the new project will still be able to deliver a great streaming service. The team also added that most of the Popcorn Time-related websites such as its blog are also going down because of DNS attacks. Popcorn Time's streaming service is gone for good, but it seems that Project Butter will be here to stay when it launches.


Achour A.,University Pierre and Marie Curie | Haddadou K.,Gandi SAS | Kervella B.,University Pierre and Marie Curie | Pujolle G.,University Pierre and Marie Curie
IEEE Communications Magazine | Year: 2012

Nowadays, wireless networks are largely deployed, and the number of smart phones has boomed, as well as data transfer via mobile phones. This has resulted in the emergence of multihoming, which allows the user to enjoy the best access for each application. The context of our work is mobile multihomed terminals in a heterogeneous environment. We propose a solution for interdomain mobility management with end-to-end service continuity of communication. For this purpose, we combine a multihoming protocol (SHIM6), which ensures a seamless network change, with the IMS architecture, which allows the establishment of multimedia sessions with quality of service. Our proposal enables a mobile terminal to change its access network seamlessly, without any application disruption. To achieve this, we implement a P-SIP inside the terminal to manage the signaling procedures. The proposed scheme can manage the interdomain mobility of a terminal in two modes: reactive and proactive. A testbed allows us to show the feasibility of our approach and the obtained gain by reducing considerably the handover delay and thus, the data loss. © 2012 IEEE.


Amamou A.,Gandi SAS | Haddadou K.,Gandi SAS | Pujolle G.,Gandi SAS
Computer Networks | Year: 2014

To achieve an efficient 'Infrastructure as a Service' (IaaS) service, virtualized data centers have to provide large-scale multi-tenancy services and to meet new requirements in terms of efficiency and availability. The main limitation of current network architectures mainly comes from inadequate design standards, such as those for Ethernet Bridging (802.1D) and Virtual LAN (802.1Q), which were not designed to handle large numbers of tenants and the resulting traffic volumes. In this paper, we present a TRILL based multi-tenant network architecture for virtualized data center networks. Through the TRILL protocol, this architecture achieves the best of both layers (bridging and routing), thus building a large-scale layer-2 network while ensuring scalability, efficiency, fault-tolerance and simplified management. In addition, this architecture integrates the VNT (Virtual Network over TRILL) solution, enabling the creation of large numbers of logical networks within a virtualized data center. The complete implementation of our solution is currently deployed and running in the GANDI public cloud infrastructure. Experimental evaluations show that VNT is able to support several hundreds of switches, thousands of physical nodes, and hundreds of thousands of tenants with only negligible overhead, all with very minimal configuration which may be completely automated. VNT additionally permits backward compatibility with both traditional layer-2 equipment as well as newer hardware supporting TRILL. © 2014 Elsevier B.V. All rights reserved.


Del Piccolo V.,GANDI SAS | Amamou A.,GANDI SAS | Dauchy W.,GANDI SAS | Haddadou K.,GANDI SAS
2015 IEEE 4th International Conference on Cloud Networking, CloudNet 2015 | Year: 2015

Cloud providers are showing an increasing interest in the TRILL protocol due to some of its capabilities such as: Virtual Machine mobility without connectivity interruption, a full non-blocking fabric, any-to-any connectivity without the use of the Spanning Tree Protocol (STP) standard, fault tolerance, and dynamic scalability. Above all, TRILL is easy to manage, plug-and-play, and does not require heavy configuration tasks to cope with virtual machine migration, changes to the network size, or physical connections. However, TRILL can not interconnect two independent networks without merging them. This is why we propose MLTP (Multi-Level TRILL Protocol) which adds a two-level hierarchy to TRILL. Each level 1 TRILL campus has independent control-plane and management while the level 2 TRILL network, the backbone, is shared between new devices, Border RBridges (BRBs) and Pseudo Gateways (PGs). This way networks can be decomposed in several smaller campuses, each one of them supervised by a different entity, thus simplifying network management. Additionally, MLTP offers a greater number of nicknames while providing tenant isolation, thus providing greater scalability. Those new functionalities are provided by MLTP without degrading throughput or latency performances as shown in our experiments. © 2015 IEEE.


Amamou A.,University Pierre and Marie Curie | Bourguiba M.,University Pierre and Marie Curie | Haddadou K.,Gandi SAS | Pujolle G.,University Pierre and Marie Curie
Proceedings - IEEE Symposium on Computers and Communications | Year: 2012

Cloud computing is an emergent paradigm that allows customers to rent infrastructure, platforms and software as a service. With resource sharing and reuse through virtualization technology, cloud environments become even more cost effective and flexible. Nevertheless, networking within virtualized cloud still presents some challenges in performance and resource allocation. In this paper, we propose DBA-VM, a Dynamic Bandwidth Allocator for Virtual Machines with regard to the established SLAs. The proposed scheme enforces the isolation between the virtual machines through the transmission bandwidth adjustment at the network I/O channel. The experimental performance evaluation shows that DBA-VM allows to the virtualized system to respect each virtual machine SLA while reducing the global physical resources (CPU and memory) consumption. © 2012 IEEE.

Loading Gandi SAS collaborators
Loading Gandi SAS collaborators