Sunday, February 3, 2013
Topic/Presenter |
---|
Full AbstractNewcomers Lunch with Old Timers (Meet and greet the NANOG community and start socializing); Sponsor: VerisignVerisign |
Full AbstractMarkley Group |
RecordingsFull AbstractSpeakers Josh Snowhorn, CyrusOne David Temkin, NANOG Program Committee Chair. |
RecordingsFull AbstractIn 2012, the communications industry faced a watershed event at the ITU's World Conference on International Telecommunications (WCIT) in Dubai. This was a treaty conference where a number of countries sought to expand a 1988 telecommunications treaty to regulate aspects of the Internet architecture, security, operations and content. Rather than bringing the world's governments together around common principles to advance the growth of telecommunications, the conference ended with countries more divided than ever on a host of important Internet issues. So, what happened at WCIT and what can the Internet community learn from the discussions? While many of us did not believe that the treaty should extend to the Internet, it is also true that countries brought important policy issues to the table that deserve ongoing discussion and dialogue. Ms. Wentworth's talk will bring the NANOG community up-to-date on what happened at WCIT, outline some possible lessons learned, and open a discussion with NANOG about the post-WCIT policy environment and what it means for operators. Speakers |
RecordingsFull AbstractThis panel will focus on the operational challenges of maintaining datacenter uptime during a major weather event like Hurricane Sandy. The panelists, all senior data center operational executives, will discuss the complexities of handling complex datacenter operations during dangerous and unpredictable weather conditions. Datacenter siting strategies will also be discussed, as well as application/datacenter interaction, especially in non-steady state conditions. The format for this panel will be a combination of panel discussion on best practices for data centers in disaster conditions, along with audience participation and questions. This panel is a must-see for network architects planning their data center strategies. Speakers Neil Crowley, Internap Scott A. Davis, DuPont Fabros Technology Panelist - Michael J. Parks, Datapipe Panelist - Michael Poleshuk, Equinix |
Monday, February 4, 2013
Topic/Presenter |
---|
Full AbstractSpeakers |
Recordings |
|
Full AbstractThe session will include a brief overview of OpenFlow's origin and the role of the Open Networking Foundation, a description of the OpenFlow protocol/API, and hands-on exercises that include an the use of an OpenFlow controller, diagnostic tools, and FlowVisor. Speakers |
RecordingsFull AbstractThe DNS 101 tutorial presents an in depth overview of the domain name system (DNS) and it's operation. We examine the core protocol, it's usage in theory and practice as well as operational examples to help illustrate the most pertinent DNS deployment considerations in real-world networks. This tutorial is for those network operators who are new to large-scale DNS administration or those who need to keep their rusty DNS knowlege and skills current. Speakers |
Full AbstractWhile IPv6 is a layer 3 protocol, major changes between versions 4 and 6 are affecting networks and applications at much higher in the stack. While it would be nice to fantasize about a world where we, as network architects, could do layer 3 and let someone else deal with other IPv6 implementation issues, we may be the best suited to identify and address these issues. End hosts might be making routing decisions and routing announcements that we don't expect. Learn what the issues are and how to work with system groups to solve these issues. Speakers |
RecordingsFull AbstractEverything you ever wanted to know about optical networking but were afraid to ask. Example topics include: * How fiber works (the basics, fiber types and limitations, etc) * Working with optics (choosing the right type, designing optical networks, etc) * Optical power (understanding dBm, loss, using light meters, etc) * DWDM (how it works, muxes, oadms, amps, etc) * Dispersion (what is it, why do we care, how do we fix it) * Optical Myths (can I hurt myself looking into fiber, can I overload my optic) Speakers |
Full AbstractMPLS P routers have trouble dealing with ethernet frames destined for a MAC address starting with a 4 or a 6. Isn't that absolutely fabulous? MAC addresses starting with a 4 or a 6 are not treated as equal citizens! Symptoms you might encounter when dealing with a L2 pseudo-wire or a L3VPN service include re-ordering, congested links or plain dropping of frames. Are you using remote peering? Is your enterprise relying on an out-sourced L3 backbone? Do you extend VLANs from one datacenter to another? YOU MIGHT BE AT RISK. A well-kept secret of router vendors, we illustrate this problem so the common operator can fully understand what is happening, why, and some thoughts on how to mitigate and fix it. Speakers Jeff Wheeler, Inconcepts. |
|
RecordingsFull AbstractHow to localize packet loss events in a large and complex network. We created a system to localize a node or a link causing end to end packet loss in a deterministic way. Speakers Nicolas Guilbaud, Google |
RecordingsFull AbstractYou're already deploying IPv6, but other parts of the Internet are still afraid of what it might cost them. Whether "they" are an ISP, data center, hosting company, or enterprise, take a look at how much it will cost to deploy IPv6, and what it will cost to operate a dual-stack network. Speakers |
Tuesday, February 5, 2013
Topic/Presenter |
---|
Full AbstractThe network operator security track explores the latest in current network security threats, defenses and research. This session will NOT be recorded. Contact the track coordinator to reserve some time to lead a discussion or present a topic of interest. Speakers |
RecordingsFull AbstractOpening and Announcements  John Curran, ARIN President and CEO ARIN-2012-2: IPv6 Subsequent Allocations Utilization Requirement  Rob Seastrom, ARIN Advisory Council ARIN-prop-182 Update Residential Customer Definition to not exclude wireless as Residential Service  Chris Grundemann, ARIN Advisory Council ARIN-2013-1 (ARIN-prop-183) Section 8.4 Transfer enhancement  Rob Seastrom, ARIN Advisory Council Open Microphone  John Curran, ARIN President and CEO Speakers |
Full Abstract6Connect |
|
RecordingsFull AbstractAn in depth analysis of internet security trends derived from Tier 1 and Tier 2 Internet Service Providers, Large Enterprises and other companies. Analysis sheds light on current threats, concerns, and trends focused on a range of network services including Internet Data Centers, Mobile Networks, and IPv6. Anonymized data from global companies with a view of over one quarter of all internet traffic. Continuous monitoring and other data sources provide insight and actionable intelligence on global DDoS activity and trends regarding botnets, malware, IPv6, and global network events which threaten Internet infrastructure and enterprise network services. Speakers |
RecordingsFull AbstractPanelists will present various Internet measurement perspectives on the impact of Hurricane Sandy on Internet connectivity: local, regional, and global. Through various technical lenses (BGP routing, active traceroute-based measurement), we will present a timeline of impairment and recovery, as well as specific provider impacts. Speakers |
Full AbstractCyrusOne |
Full AbstractFor years, the U.S.-to-Latin America submarine cable market has been characterized by high prices and limited competition. Similarly, the cost of IP transit in South America remains many times the price of IPT in North America in Europe. The past two years have seen the announcement of a spate of new cables that hold the promise of alleviating the situation for buyers and users of international bandwidth. Investors of various stripes have proposed a bewildering array of new cables that, if built, could transform the market in Latin America. The presentation will share results from primary research findings addressing the following questions: * How quickly has bandwidth demand increased to/from Latin America, and which cities serve as the primary hubs? * How much do transport and IPT services cost to and within Latin America? * Which cables have been proposed for the region? Which have actually achieved funding? * What are reasonable expectations for market prices when new cables are built? Speakers |
Full AbstractMeasuring Quality of Experience (QoE) with YouTube from users accessing from Residential ISPs is often a difficult task. First, there is the problem of a good data collection mechanism, second is the issue of being able to obtain precise information on metrics that impact QoE such as video start time, number of interruptions, duration of interruptions, etc. Due to YouTube's content distribution architecture, it is also important to be able to obtain information on how the location from where a video is served impacts QoE metrics for end users. We have developed Pytomo, a tool that emulates YouTube viewing experience by end users, through which we can collect a number of QoE metrics along with information on CDN. The tool is made available in the public domain at http://code.google.com/p/pytomo/ Using this tool, we have collected data from end users accessing YouTube from three different residential ISPs. In this talk, we will present observed QoE along with impact due to CDN. We will also discuss how such a tool can be useful to ISPs ( (it has been deployed by one ISP). This is a joint work by Parikshit Juluri*, Louis Plissonneau+, Yong Zeng*, Deep Medhi* (*University of Missouri-Kansas City, USA; +Orange Labs, France) Speakers |
Full AbstractNetworking challenges in data center and cloud environments have received significant attention by industry and standards organizations. The data center environment is dominated by the presence of software networking components (vswitches) in server hypervisors, which may outnumber by an order of magnitude the physical networking nodes. Bridging the gap between server based networking and existing network services is a significant challenge, since the ultimate goal is the design of end-to-end network services. Existing solutions rely on static routing and/or centralized routing mechanisms that cannot meet the requirements for resiliency and dynamic networking. This talk will discuss a simple approach for this problem, which combines traditional control plane and routing protocol approaches with the flexibility of SDN architectures and Openflow. The mechanism relies on maintaining an Openflow interface to hypervisors and utilizing existing routing mechanisms when a scaled out data center deployment is required to federate a number of SDN controllers, or to interoperate with traditional network services without the need for dedicated gateways or complex OSS integrations. Speakers |
|
RecordingsFull AbstractBIRD was presented on NANOG 48, last time. Since that time, many new features have been added like ROA support, extended communities filtering, template system and improved filtering capabilities etc. The presentation will show examples of real implementation in IXPs. Speakers |
RecordingsFull AbstractNetwork World recently argued that "it will be difficult for Internet policymakers, engineers and the user community at large to tell how the upgrade to IPv6 is progressing because no one has accurate or comprehensive statistics about how much Internet traffic is IPv6 vs. IPv4." We have been seeking to address this concern by collecting broad visibility, multi-perspective data on adoption. In this talk, we present a snapshot of global IPv6 deployment via ten measurements taken from seven large datasets. Together, our measures seek to provide a deep and broad view of the process by characterizing IPv6 addressing, transition technologies, DNS readiness, host readiness, application mix, and traffic volume. Speakers |
RecordingsFull AbstractJohn Curran, President and CEO of ARIN, will give the latest news about ARIN including the status of IPv4 depletion, IPv4 address transfers, and IPv6 uptake in the ARIN region. He will also discuss the continued development and integration of ARIN Online, RPKI, and ARIN’s participation in Internet governance forums. During the Q&A portion of the ARIN Update today given at NANOG 57, referenced was made of some letters sent and received with regards to legacy addresses. here is it: https://www.arin.net/resources/legacy/index.html The particular question is in the FAQ section at the bottom (question #18); the links to the documents (in pdf form) are in the response text for that question. Speakers |
|
Full AbstractSpeakers |
|
Wednesday, February 6, 2013
Topic/Presenter |
---|
|
Full AbstractAkamai |
Full AbstractPower Breakfast - Start your day with a vast selection of hot/cold food for all tastes! |
Full AbstractNetflow/sFlow are a critical tools for network infrastructure monitoring and capacity planning. However, collecting and analyzing flow and other telemetry data for large scale networks could be a challenging task due to vast volume of data that has to be stored and joined with large external datasets for analysis. Commonly, relational databases have been used to store flow data and run OLAP-style analysis queries, but with very large data-sets such systems become prohibitively complex to operate. In this presentation we are sharing experience on building a scalable system for various telemetry data collection leveraging horizontally-scalable software load-balancers and using Map-Reduce framework for data mining. We demonstrate how a Map-Reduce type system could be used to store both flow-type and BGP data and build reports joining large data-sets together. Though our system is built using in-house tools, this approach could be reproduced using open-source software. Speakers |
Full AbstractThe rate of growth of the global routing (RIBs) and forwarding tables (FIBs) has been increasing exponentially for the past several years. The growth of these is due to fac- tors such as multihoming, traffic engineering and prefix deaggregation. IPv6 and future architectures that route on name prefixes could lead to an even faster table growth. The FIB growth is a more pressing concern since FIBs are stored on fast but expensive card memory, costing thousand of dollars per GB. In this paper, we introduce a FIB caching system that in- troduces a novel hole filling algorith to overcome the cache hiding problem. We use LRU caching to substantially re- duce the number of FIB entries required to serve the entire routing table from the current 350K to 5K while achieving over 99% hit rates. We investigate trends over a period of four years and show that the required cache size remains virtually constant. For cache misses, we estimate the buffer requirements to hold packets while the new route is loaded into the cache and show that only a few hundred packets need to be buffered. Further, we investigate the robustness of our caching solution, where an adversary attempts to poison the cache. Finally, we investigate the impact of dynamic BGP updates on the performace of our caching solution. Speakers |
RecordingsFull AbstractSpeakers |
Full AbstractNetwork Operators and Service Providers are facing increased regulatory and consumer pressure to optimize the power consumption and efficiency of consumer network devices and the network as a whole. In addition to device manufacturers addressing these power consumption efficiency trends, network monitoring applications need to evolve in the back office to monitor and manage the energy and power attributes of the network, on an elastic dynamic basis. This presentation proposes an architectural approach to addressing energy management of the network by highlighting a unified approach to energy management metrics definition and a standardized methodology for efficient collection of those metrics from the network. Speakers |
|
RecordingsFull AbstractBy tweaking the BGP configurations, ASes are able to express their interdomain strategies and deploy the above-mentioned policies. However, due to the complex interactions between networks, at times only defining a routing policy does not also guarantee its effectiveness in the Internet. For examples, sometimes, BGP routes end up being announced to entities to which the routing policies specifically state that they should not be. And it may also happen that, by default, all the routes are propagated to all BGP neighbours, once they have been accepted, disregarding of the origin’s strategies. Consequently, the network operators need to complement the view of the routing system from inside their network with multiple external views, in order to understand how their policies interact with the operations of the rest of entities in the interdomain and if the implemented policies are producing the intended or expected result. Speakers |