[[PageOutline]] = I&M WG meeting at GEC7: Agenda and Notes = Wednesday, March 17, 2010, 3:30pm - 5:30pm [[BR]] Room: Presidents 2 [[BR]] == Introductions == 3:30pm All [[BR]] == Major WG issues and goals == 3:35pm Paul Barford (University of Wisconsin) [[BR]] [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/barford%20%20gec7_imwg.pdf slides] (15min) [[BR]] Goals: [[BR]] + An infrastructure for gathering, analyzing and archiving measurements in GENI [[BR]] - Support for broad range of experiments [[BR]] - Tight integration with control framework(s) [[BR]] - Efficient, easy to use, broadly deployed, diverse capabilities, secure, privacy-aware, etc. [[BR]] + WG facilitates communication and coordination between I & M projects [[BR]] - Toward reaching overall goals quickly and effectively [[BR]] - Includes architectures, specifications, partnerships [[BR]] Issues:[[BR]] + Architectural (v0.1 on wiki)[[BR]] - Use cases[[BR]] - Schema[[BR]] - Sensors, components and protocols[[BR]] - UI and integration[[BR]] - Authentication and privacy[[BR]] - torage and analysis[[BR]] + Practical[[BR]] - Test and evaluation[[BR]] - Deployment, configuration and support[[BR]] == Project Reviews == Each of the following speakers was asked to include in their talk a review of how they address the following GENI I&M architecture priority topics: [[BR]] + Common terminology; best granularity of functions [[BR]] + Measurement data schema; common after MPs, before and within MCs, MDAs; what is included in meta-data? [[BR]] + Measurement plane; options; expect nodes with 3 or 2 NICs [[BR]] === Instrumentation Tools Project (1642) === 3:50pm Jim Griffioen (University of Kentucky) [[BR]] [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/griffioen%20%20instools_imwg_gec7.pdf slides] [[BR]] Goals:[[BR]] + Integrate Univ of Kentucky Emulab into ProtoGENI (completed in year 1)[[BR]] + Reimplement Univ of Kentucky Edulab instrumentation and measurement tools to work in the ProtoGENI environment.[[BR]] + Support automatic generation of instrumentation and measurement infrastructure on a per-slice basis.[[BR]] Structure:[[BR]] + Measurement Points[[BR]] + Measurement Controller[[BR]] Functional Components[[BR]] 1. Setup: deploy and initialize topology-specific software and services[[BR]] 2. Capture: capture measurement data[[BR]] 3. Collection: move data to processing/storage environments[[BR]] 4. Storage: store data on a temporary, short term, long term, and archival basis[[BR]] 5. Processing: filter, convert, aggregate, summarize, etc., data[[BR]] 6. Presentation: present data to users in meaningful ways[[BR]] 7. Access Protection: protect resources and data[[BR]] 8. Measurement Control: Dynamically control the above components[[BR]] Note: Conventional network management solutions exist for 2. through 8.[[BR]] Mapping functions to proposed services:[[BR]] 1. Setup – MO?[[BR]] 2. Capture - MP[[BR]] 3. Control – MO?[[BR]] 4. Collection - MC[[BR]] 5. Storage - MDA[[BR]] 6. Processing - MAP[[BR]] 7. Access Control – MO?[[BR]] 8. Presentation - MAP[[BR]] Measurement data schema:[[BR]] + Data-specific formats will exist at multiple times and places but should be invisible to the generic measurement infrastructure[[BR]] + Should leverage expertise of those with experience in this area (DatCat, PerfSONAR, etc).[[BR]] Measurement plane:[[BR]] + Virtual interfaces abound in GENI. [[BR]] - Unknown how virtual interfaces map to physical NICs.[[BR]] + Shared paths are OK, but require QoS.[[BR]] - Setting up QoS planes is challenging; particularly if planes are per-slice.[[BR]] === OMF/OML Project (1660) === 4:05pm Max Ott (NICTA) [[BR]] [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/ott%20%20OML%20-%20GEC7%20-%20march.pdf slides] [[BR]] Goals:[[BR]] • All experiment output in one place[[BR]] • Capturing everything – most importantly meta data[[BR]] • Separation of concerns[[BR]] – Instrumenting[[BR]] – Collecting[[BR]] • Minimizing measurement collection overhead[[BR]] – Time[[BR]] – Traffic interference[[BR]] • Support for steerable experiments[[BR]] – Access to data in different places[[BR]] Concepts:[[BR]] + MPoints[[BR]] + Filters[[BR]] – Stddev[[BR]] – Average[[BR]] – First[[BR]] – Histogram[[BR]] + Processing/Caching[[BR]] + Steering or Feedback[[BR]] + MStreams[[BR]] + Dynamic Schema[[BR]] + Visualization[[BR]] Supported applications:[[BR]] • Traffic Generation/Measurements[[BR]] – OTG … Traffic Generator[[BR]] – Iperf[[BR]] • Monitoring[[BR]] – Libtrace[[BR]] – Libsigar[[BR]] – Spectrum Analyzer[[BR]] – GPS[[BR]] – (Weather)[[BR]] • Components[[BR]] – TinyOS/Motes[[BR]] – (GnuRadio)[[BR]] === LAMP using perfSONAR (1788) === 4:20pm Guilherme Fernandes (for Martin Swany) (University of Delaware) [[BR]] [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/fernandes%20%20swany%20%20GEC7-LAMP.ppt slides] [[BR]] + perfSONAR is a multi-domain performance monitoring framework, which defines a set of protocol standards for sharing data between measurement and monitoring systems [[BR]] Architecture: [[BR]] + Interoperable network measurement middleware designed as a Service Oriented Architecture (SOA): [[BR]] - Components are Web Services (WS) based [[BR]] + Several unique components and design considerations, all of which operate in a cooperative yet independent manner [[BR]] - Each functionality is separated into a specific function [[BR]] - Clients and servers interact through scripted, XML Based protocols [[BR]] - Measurement data is encoded in expressive XML formats [[BR]] Components: [[BR]] + Infrastructure [[BR]] - Lookup Service [[BR]] - Topology Service [[BR]] - Authentication Service [[BR]] + Services [[BR]] - Measurement Point (MP) Service [[BR]] - Measurement Archive (MA) Service [[BR]] - Resource protector [[BR]] + Analysis and visualization [[BR]] Open protocols and schema: [[BR]] + Base network measurement schema[[BR]] - OGF Network Measurement Working Group[[BR]] + Topology Schema[[BR]] - OGF Network Markup Language (NML-)WG[[BR]] - Includes Topology Network ID[[BR]] + perfSONAR Protocol Documents[[BR]] - OGF Network Measurement and Control (NMC-)WG[[BR]] Base network measurement schema:[[BR]] + Measurement Data, a set of measurement events that have some value or values at a particular time[[BR]] + Measurement Metadata, the details about the set of measurement data[[BR]] Measurement metadata:[[BR]] + Subject (Noun)[[BR]] - The measured/tested entity (who)[[BR]] - E.g. A pair of hosts (end-point-pair), or a Layer 3 interface[[BR]] + EventType (Verb)[[BR]] - What type of measurement, value, or event occurred[[BR]] - Characteristic, tool output, or generic event[[BR]] - E.g. latency, bandwidth, utilization, or simply iperf[[BR]] + Parameters (Adjectives and Adverbs)[[BR]] - How, or under what conditions, did this event occur?[[BR]] - E.g. buffer sizes used, TCP vs ICMP packets[[BR]] + Key[[BR]] - Shortcut substituted in place of previous three items[[BR]] - No predefined format[[BR]] Measurement data: + Datum: The actual result (values) of measurement. [[BR]] - Can contain time (e.g. a Time element or attribute).[[BR]] - Existence of an event might point to the case where there no additional value[[BR]] - As in “Link up/down” or threshold events[[BR]] + Time: Representation of a time stamp or time range in a specified format.[[BR]] - Must be extensible since even agreement about the right structure is not easy, e.g. UNIX timestamp vs NTP time[[BR]] Schema namespaces and extensibility:[[BR]] + A namespace: http://ggf.org/ns/nmwg/base/2.0/[[BR]] - MAY NOT be a URL[[BR]] + We encode the measurement/event type in the namespace (and as a standalone element)[[BR]] + We use Data and Metadata elements and vary the namespaces of the specific elements[[BR]] + Extensibility achieved through hierarchy with delegation[[BR]] - Similar to OIDs in the IETF management world[[BR]] + The NM-WG has a hierarchy of network characteristics[[BR]] - Good starting point[[BR]] - E.g. http://ggf.org/ns/nmwg/characteristic/utilization/2.0[[BR]] - E.g. http://ggf.org/ns/nmwg/characteristic/bandwidth/achievable/2.0[[BR]] + However, not all tools are cleanly mapped onto the Characteristic space[[BR]] - Often a matter of some debate[[BR]] + Organization-rooted tools namespace addresses this[[BR]] - Easy to add new tools in organization-specific namespaces[[BR]] - E.g. http://ggf.org/ns/nmwg/tools/nuttcp/2.0[[BR]] Topology Schema: [[BR]] + Topology schema grew from network measurement description [[BR]] - Reusable “Subject” elements for common cases [[BR]] - Also reduces redundancy [[BR]] - Relationships between measurement Subjects [[BR]] + Structured by layers and the same elements recurring there (Base, L2, L3, L4) [[BR]] - networks as graphs[[BR]] + Elements:[[BR]] - Domain[[BR]] - Node[[BR]] - Port[[BR]] - Link[[BR]] - Network[[BR]] - Path[[BR]] - Service[[BR]] + Varied by namespaces (extensibility)[[BR]] - Reuse visualization logic, etc.[[BR]] - Validate layer- or technology-specific attributes[[BR]] + Used by perfSONAR, IDC Protocol (ION, OSCARS, AutoBAHN), Phoebus[[BR]] - Currently calling it the UNIS Topology Schema [[BR]] + OGF NML-WG to unify NDL and UNIS Topology schema[[BR]] - Happening as we speak at OGF28[[BR]] LAMP objectives:[[BR]] + Collaborate on defining a common but extensible format for data storage and exchange for GENI I&M systems[[BR]] - Use perfSONAR NM-WG schema as starting point[[BR]] - Identify new characteristics/tools namespaces[[BR]] + Develop a representation of GENI topology to be used to describe measurements and experiment configuration[[BR]] - UNIS topology schema can be easily extended [[BR]] + Collaborate with related GENI measurement and security projects on a common GENI I&M architecture[[BR]] - The new GENI I&M Arch. Draft defines very similar services (MP, MC, MDA, MAP), and new ones (MO)[[BR]] - perfSONAR is a good starting point, not currently a final solution (for GENI); [[BR]] - Use cases have been different, but much can be reused and the framework can be extended[[BR]] === DatCat Project === 4:35pm Brad Huffaker (CAIDA, UC San Diego Supercomputing Center) [[BR]] [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/huffaker%20%20dat%20cat%20%20%20genie-gec07-201003.pdf slides] [[BR]] Goal: [[BR]] + DatCat was designed to improve data sharing by providing a unified metadata database for Internet data. [[BR]] + Make easy for users: [[BR]] - finding data sets of interest [[BR]] - adding new data sets to the catalog [[BR]] - annotating data sets in the catalog [[BR]] + DOES NOT store data [[BR]] Database scheme: [[BR]] + Collection [[BR]] - logical group of files(paper, project,...) [[BR]] + Data files [[BR]] - raw data files (traceroutes, logs dumps, ...) [[BR]] + Packages [[BR]] - downloadable files (single file, tarball, ...) [[BR]] + Locations [[BR]] - how to get the packages (URL, contact address, ...) [[BR]] Annotations: [[BR]] + Provide an extensible naming space for assigning domain specific values to files. [[BR]] + each user has their own hierarchical name space [[BR]] - passive.IPv4.packet_count [[BR]] - active.RTT_95th_percentile [[BR]] + both data contributors and general DatCat users may attach annotations [[BR]] + any user may assign “note” annotations to any object [[BR]] Metadata fields: [[BR]] + collection [[BR]] - fields: name, contents, summary, motivation, creators/primary contact/contributor, start/end time, keywords, short description/description/description URL [[BR]] - annotations: note [[BR]] + data [[BR]] - fields: name, creators/primary contact/contributor, keywords,format, file size, start/end time, duration, geographic/network location, time zone, MD5, description, creation process [[BR]] - annotations: passive.IPv4.packet_count, passive.IPv4.TCP.dst.port_count, cfg.passive.capture_len, AS_count, active.trace_count, active.RTT_10th_percentile, ..... [[BR]] + location [[BR]] - fields: package, creators, primary contact, status, download procedure, download URL, geographic/logistic location, availability [[BR]] Submission tools: [[BR]] + Perl API [[BR]] - useful for integrating into existing data management systems [[BR]] - flexible, but need to write code: [[BR]] + subcat [[BR]] - different approach (declarative) [[BR]] - describe metadata in human-friendly text files (YAML) [[BR]] - CAIDA provides tools to extract additional metadata (data-to-yaml) [[BR]] - subcat intuitively joins information together [[BR]] DatCat web portal: [[BR]] + Browse collections [[BR]] + Search collections [[BR]] + Search data [[BR]] Lessons learned: [[BR]] + file-level metadata hard [[BR]] - hard to fix errors across thousands of files [[BR]] - hard to display thousands of files [[BR]] - hard to generate [[BR]] + submission process too cumbersome for most users [[BR]] - majority of metadata is shared between files, creator, creation process, location, etc [[BR]] - many researchers are not programmers [[BR]] - researchers have limited time and motivation [[BR]] + Lots of redundant information: [[BR]] - For a single contribution, a majority of data objects have identical metadata shared across a large number of data objects. [[BR]] - could be solved by pushing subcat-type categories into the database [[BR]] + Move to stand-alone collections [[BR]] - contributors will only need to fill in the collection information [[BR]] - shorten search path from collection to locations [[BR]] + better to have lots of collections, than lots of files [[BR]] == GENI I&M Architecture == 4:55pm Harry Mussman (GPO) [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/031210%20%20IM-ARCH-GEC7Slides.pdf slides] [[BR]] [http://groups.geni.net/geni/wiki/GeniInstrumentationandMeasurementsArchitecture GENI I&M Architecture document] (15min) [[BR]] v0.1 DRAFT includes proposed I&M services and proposed configuration [[BR]] Priority topics: [[BR]] + Common terminology; best granularity of functions [[BR]] + Measurement data schema; common after MPs, before and within MCs, MDAs; what is included in meta-data? [[BR]] + Measurement Plane; options; expect nodes with 3 or 2 NICs [[BR]] == Next Steps for WG == 5:10pm Bruce Maggs (Duke University) [[BR]] == Wrap-up, review of action items and issues for planary == 5:25pm Harry Mussman (GPO) [[BR]] [http://groups.geni.net/geni/attachment/wiki/Gec7InstMeasWGAgenda/031710%20WGInstAndMeas_Report.ppt slides] [[BR]] 5:30pm Adjourn [[BR]] 6:30pm BoF dinner, organized by Harry Mussman, location [http://www.ghgrestaurants.com/parizade/parizademaster.html Parizade Restaurant][[BR]]