ONOS Goldeneye Release Content

Community - New collaborators

• Announced new collaborators: Calix
• Unannounced new collaborators that have been approved but haven’t been communicated yet: Argela, IPT, SDNLAB, Villa-Tech

R&E Deployments

• Contributions from ON.Lab, AmLight, AARNET, Internet2, GEANT, IEEE
• Enable AmLight to change cost of the links in their network
• Castor has been contributed back by AARNET to the ONOS main repo
• Initial steps to integrate L2 and L3 apps (specifically VPLS and SDN-IP) in ONOS
• Announced ONOS integration into an international industrial federated testbed

Automated testing, Build environment

• Contributions from ON.Lab, Ciena, Huawei, Criterion

• Phased in "Chaos Monkey" for longevity test

• Built test suite on dynamic clustering

• Included testing intent with FlowObjectives

• BGPLS functional test case

• IPv6 functional test case

• Added VLAN selectors and treatments  to Intent functionality tests

• Moved to Buck in the build environment

Applications

Contributions from ON.Lab, Huawei, POSTECH, CREATE-NET, Politecnico of Turin, Tata (TCS), BUPT

• ICONA enhancements - for example, ICONA to BGP interface
• Control Plane Manager (CPMan)
  • Control Plane Manager adds management capability to the control plane enhancing HA/reliability
  • Added REST API for exposing control metrics (e.g., system metrics + control message)
  • Enhanced MetricsDatabase to support multi-resources
  • Enhanced metric reporters in a way to notify all metric reporters when metric registry has been changed
  • Support for querying control metrics of remote node
  • Implemented Web GUI using chart model to visualize control metrics in bar/line charts
• InfluxDB Report and Query App
  • This application adds capability to report and retrieve various control metrics stored in MetricsService to/from InfluxDB
  • By querying InfluxDB, we can easily obtain the control metrics from all ONOS instances
  • Added a way to integrate InfluxDB and Grafana to visualize control metrics
• Ganglia Report and Query App
  • This application adds capability to report various control metrics stored in MetricsService to ganglia monitoring server
• Chart Model
  • Chart model provides a way to visualize a set of data points with line or bar chart without knowing the details
  • CPMan makes use of chart model to plot control message stats in both line chart and bar chart
• Network Troubleshooting App
  • The Loop Detect interface and the Black Hole Detect interface are integrated into one interface. Network troubles can be detected by calling this unique interface.
  • Improved code structure and function partitions.
  • The application supports Remote Procedure Call, helping to prevent blocking when multiple applications call the network trouble shooting function.
  • Improved the javadoc making the API comments generated by swagger more friendly.
• SDN-IP App
  • Runtime configuration update support.
  • Remove intents using App ID when App get deactivated.
  • Reactive-routing uses new config subsystem of ONOS. Which means, there is no need to use son-ip.json any more.
  • Introducing support for multi vlans for SDN-IP
• Advanced Yang constructs implementation
  • Include, Import, Grouping/Uses, Choice, RPC, Notification, Augment, Sub-module, Inter/Intra file linking, all built-in and derived data types.
• SFC
  • Enhanced interfaces with classifier, SFF, SFC Proxy
  • Maintain SFC statistics
  • GUI for viewing SFC resource statistics
  • Tie-break logic implementation at classifier
  • Interface with SFC proxy
  • Added new criterion and actions to support NSH as per official OVS patch
• BGP
  • Added capability to support BGP flow specification and route policy distribution
  • Added REST API for configuring flow specification and route policy distribution
  • Protocol enhancement to support flow specification
  • Protocol enhancement to support route policy distribution
  • code reorganized to support complex flow classes
  • Runtime configuration support

• PCE/PCECC
  • Computation of optimal E2E paths (PCE) for MPLS TE tunnels within a PCE domain based on network topology information and computational constraints.
  • Extension to PCE to act as a centralized controller (PCECC) where in the PCECC of controller takes care of label distribution and thus avoid signalling protocols like RSVP-TE, LDP etc.
  • Features include:
    • PCEP protocol enhancement to support stateful PCE/PCE Initiated LSP/PCECC.
    • CSPF path computation in ONOS PCE
    • PCE Global path optimization
    • WebUI enhancement for managing PCE tunnels
    • Label management for PCECC (PCE as Centralized Controller)
    • Support PCECC-Basic using PCEP protocol
    • Support PCECC-SR (Segment Routing) using PCEP protocol

Northbound

• Contributions from ON.Lab, Ciena, POSTECH, GEANT,  CREATE-NET, Politecnico of Turin, Tata (TCS)
• Northbound API support for ICONA enhancements
• Intent subsystem now integrated with Flow Objectives subsystem
• Intent-based implementation of virtual network provider. It is really not suitable for production yet, but it does validate the mechanics of virtual-to-physical mapping and isolation
• Added CLI & REST API for managing the inventory of the virtual networks
• Develop STC scenarios for validating the v-net functionality and to help further test the functionality during development
• Initial work on on adding various ONOS services atop virtual networks, starting with device, link, host
• REST API improvements
  • Supports flow table pipeline
  • Supports encode/decode meter/group/sequence id
  • Bumped up Jersey from 1.x. to 2.x
  • Now can specify application identifier when inserting flow rules
• Web UI stylesheets refreshed to use look and feel consistent with project website.
• Started to refactor the Web UI Topology View to be "Region Aware" -- to support enhanced interaction / navigation of the topology
• Introduction of CLI command "device-interfaces" that lists interface information for legacy devices

Core

• Contributions from ON.Lab, Fujitsu, ETRI, NEC
• ~6x improvement in throughput for distributed strongly consistent operations
• Performance and stability enhancements for Resource Store by optimizing the manner in which we represent individual resources in data store
• ECMap enhancement to periodically purge old tombstones (deleted item placeholders). Improves the stability of long running tests
• Refactored optical information model into a loadable application using the projection mechanism' making the ONOS core easier to scale and more modular. It is also a blueprint for integrating other technologies such as microwave support.
• Improvement in serialization mechanism to detect object registration id conflict detection.
• Created a core unicast route service which provides a centralized route store in ONOS. This helps facilitate the growing number of routing apps and use cases on ONOS by allowing more collaboration between apps and greater API stability.
• Performance improvements to Adaptable Flow Statistics subsystem
• Refactor of core Alarm subsystem:
  
  • polling alarm provider that uses drivers to query alarms
  • distributed store for alarms
  • AlarmService with listener mechanism
  • unit tests

Southbound

• Contributions from ON.Lab, GEANT, Huawei
• Refactor SNMP implementation:
  • correct division of elements in provider, protocol and drivers.
  • SNMP subcontroller.
  • drivers for BTI, Lumentum
  • unit tests
• Lumentum Alarm Driver
• Initial support for the reference P4 Behavioral Model (BMv2) software switch
• Mastership for NETCONF devices
• Introduction of NETCONF Controller component configuration for "netconfReplyTimeout" setting the timeout for NETCONF synchronous replies
• SB interface for retrieval of interfaces configuration from legacy devices and its implementation in Cisco IOS NETCONF driver
• Handling of multiple VLAN-IDs for trunking interfaces on legacy devices and its implementation in Cisco IOS NETCONF driver
• Removal of VLAN-ID when removing trunk/access mode from legacy device interface
• YANG Southbound tool for drivers
  • yang to NETCONF XML translation