IAB Business Meeting: 2019-11-18, Singapore
- Jari Arkko
- Stephen Farrell
- Wes Hardaker
- Ted Hardie (IAB Chair)
- Christian Huitema
- Olaf Kolkman (Internet Society)
- Zhenbin Li
- Cindy Morgan (IAB Executive Administrative Manager)
- Erik Nordmark
- Mark Nottingham
- Karen O’Donoghue (Internet Society Liaison)
- Colin Perkins (IRTF Chair)
- Melinda Shore
- Jeff Tantsura
- Martin Thomson
- Brian Trammell
- Alissa Cooper (IETF Chair)
- Mirja Kühlewind (IESG Liaison)
- Marie-Jose Montpetit, COINRG Chair
- Jeffrey He, COINRG Chair
- Eve Schooler, COINRG Chair
1. IRTF RG Review: Computing in the Network Proposed Research Group
Marie-Jose Montpetit, Jeffrey He, and Eve Schooler joined the IAB to give an update on the activities of the Computing in the Network Proposed Research Group (COINRG).
The goals of the RG are to create a research community around COIN principles to improve network performance for services, applications & users, and to participate in the evolution toward in-network route-compute-store network elements. The scope includes architectures, protocols, and real-world use cases and applications. The focus is on the core-to-edge compute continuum.
The community is a good mix of computer scientists, applications developers, and implementers in academia and industry R&D. The community includes international research initiatives currently in South America, Europe, Canada, China, and southeast Asia.
COINRG is looking at the following implications/questions:
- Where in the HW/FW/SW stack will the Compute expect to be performed?
- For truly improved performance, Network functionality needs to be designed to play well not only with Compute, but also with Storage.
- Developers/users don’t want to have to specify where exactly COIN will run in the Cloud-to-Edge computing continuum
- If an abundance of Data originates at Edge devices and flows upstream, Edge aggregator nodes may need to perform Compute to reduce resource usage and/or to manage data implosion
Research objectives include systematically looking at different instantiations of COIN in the cloud-to-edge-to-device computing continuum:
- Are there common principles, abstractions, assumptions and mechanisms that can be applied across this range of different types of computing/networking applications/approaches?
- What are best practices and relevant considerations for computing/networking systems, in particular with respect to previous efforts in active networking?
- For a network with non-forwarding functions, at what level of abstraction must the programmable data plane live?
- What is the impact of in-network functions on end-to-end transport protocols and security?
- What are the incentives for the network to add new computing & storage capabilities and resources in an open eco-system?
COINRG has several research-driven I-Ds in progress:
- draft-montpetit-coin-xr, In Network Computing Enablers for Extended Reality
- draft-mcbride-edge-data-discovery-overview, Edge Data Discovery for COIN
- draft-he-coin-managed-networks, In-Network Computing for Managed Networks: Use Cases & Research Challenges
- draft-sarathchandra-coin-appcentres, In-Network Computing for App-Centric Micro-Services
- draft-kunze-coin-industrial-use-cases, ndustrial Use Cases for In-Network Computing
- draft-liu-coinrg-requirement, Requirement of Computing in network
- draft-kutscher-coinrg-dir, Directions for Computing in the Network
- draft-kunze-coinrg-transport-issues, Transport Protocol Issues of In-Network Computing Systems
Christian Huitema noted that the current list of drafts does not include much on trust, security, or privacy. Eve Schooler agreed that they need more work there; Marie-Jose Montpetit said that they will make a point of recruiting more people to work on those topics.
The next steps for COINRG include:
- Articulating problems & challenges, potential “gains”
- Studying Commonality across Cloud-to-Edge computing continuum
- Identifying (more) real use cases and their requirements
- Comprehending Programmability models
- Soliciting proposals for architecture, protocols, eco-system solutions
- Developing group Hackathon project(s) that showcase route-compute-store convergence and a range of in-network computing
- Continuing to refine charter to scope down!
- Continuing to foster informal dialog w/related groups: e.g.,
- IETF WGs/RGs: T2T, DIN, ICN, DetNet, CoRE, DNSSD, TEEP, RATS
- External consortia: ETSI MEC, LFEdge/EdgeX, ECC, IIC/OpenFog
Christian Huitema asked what the relationship is between the COIN architecture and the move towards encrypted transport. Marie-Jose Montpetit replied that it is likely to be an issue; it is on their list of things to address.
2. Potential 2021 Retreat in Dagstuhl
Stephen Farrell asked if the IAB should consider trying to plan a 2021 retreat in Dagstuhl, given the long lead time needed to reserve space there. The IAB agreed that while Dagstuhl would be a good location for an IAB retreat, the uncertainty about the availability of those who would be on the IAB in 2021 made planning a retreat that far in advance too challenging.
3. Potential Program about Internet resiliency
Stephen Farrell reported that he had a conversation with Terry Manderson at IETF 105 about a possible IAB Program on Internet resiliency.
Christian Huitema noted that this is related to the Internet centralization issues.
Stephen Farrell, Wes Hardaker, Christian Huitema, and Brian Trammell will draft text for a description of an IAB Program on Internet resiliency. Brian will set up a sub in the IAB Github repo for this work.
4. ICANN GWG Appointments
The IAB agreed to run a shortened version of the normal IAB appointment process in order to appoint two members to the Root Server System Governance Working Group by 10 January 2020. Cindy Morgan and Jari Arkko will draft text for the call for nominations and send it out before the IETF 106 plenary.