Decentralized Internet Infrastructure Research Group (dinrg)
- IETF 105, Montreal
- 24 July 2019, Van Horne Room
- Note taker: Leandro Navarro
Introduction about the WG
Leandro Navarro, “MeshDApp: Blockchain-enabled Crowdsourced Internet Access Platform for Mesh Networks”
- Routing of traffic and value flows to compensate investment in network infrastructure, maintenance and other costs, in a decentralized/crowdsourced network, using a permissioned blockchain using a private Ethereum PoA network, one per mesh network. Uses an Oracle to * connect the traffic monitoring/accounting system and the economic ledger system.
- Smart contracts can help to ensure future transactions happen according to pre-established conditions.
- The ledger records transactions (fungible token transfers) across participants (either consumers and providers) and records the ownership (mining, registration) of devices to participants (non fungible tokens).
Q: blockchain, only controlled entitiy can write to it?
A: you can see what’s happening in the network, and admitted node will run the smart contrract
Q: most of the work is done by Oracle, not by Blockchain. You could also just publish everything in Blockchain. What’s the added value of blockchain?
A: managing accounting and transferring is some effort, idea is to automate that manual process of managing money.
Q: does the blockchain contain info of who is sending info where? Privacy?
A: blockchain aggregates info about traffic (volume). Statistics per node. Want to increase privacy to make sure connection info is anonymous (future work).
Shen Yan, “A Blockchain based Testbed for BGP Verification”
- Distributed Internet Infrastructure
- Related to BGP
- BGP lacks verifiability of announcement messages: hijack (origin and path), …
- RPKI -> decentralized trust model, but sitll routing issues
- A testbed with Hyperledger Fabric (HLF), consensus system, in each AS 1 HLF node
- In each Fabric node: IP and ASN onwership, ROA (IP->ASN), ASNeighbor -> can do validation
- Operation: BGP Origin verification, origin verification, route leak protection
- Testbed: HLF, 6 organizations, planning phase, open
- Q: Suggestion about storing in blockchain a summary/hash of BGP info, as blockchain grows
- Q: Comment signatures, etc.
- Q: Who has authority to claim ownership about network ranges, initial determine claims, a relation in the ledger is real, unique, authentic
- A: Original information. Use new IP assignments (uses smart contract to assign), consensus but … check validity of the range. They use info from existing network.
- Q: Attributing ownership to business ownership, not done properly for diverse reasons.
- A: Registering information in the ledger allows to access benefits. Want to provide ways to BGP verification
- Q: Whos is the source of authority. Need for a political decision to decentralize. Need for a log. We built PKI X509 wrong tech, just a log of allocations with a merkle tree, etc. In fact, centralising the current system even more. Anything that involves SC is not appropriate for this problem.
- Q: Clarification about Consensus
- Q: Centralized vs hierarchical vs decentralized. You’ve built a ledger but not solved …
- Q: Take network prefix and get it assigned, but the requester may not have the right to get it …
- Q: Allocation prefixes to owners + record payment
- Q: Representation of , magic , public ledger is the public proposition
- Q: About recovering keys, comments about private/public ledger
- Q: SC to lease a subrange? SC are not so bad
- Chairs: Need to take the critique seriously, continue discussion on the mailing list.
Lixia Zhang, “Decentralization: from the ground up”
- DLedger: An IoT-Friendly Private Distributed Ledger System Based on DAG
- Decentralization has to be from the ground up.
- Why blockchain? because it has no controller
- Not really if bitcoin, only the big players can win in PoW = proof of resources (centralization?)
- Yes if blockchain is e.g. Hyperledger, no anonymous participants but trusted - permissioned.
- DLedger: a private distributed ledger system: all entities form a p2p network, each peer appends new records, …
- Security through publicity (PSKI Usenix Hotsec 2006): acting in public provides info and semantics to construct meaningul security mechanisms.
- Importance of a unique namespace system, a hierarchy is needed, like in DNS.
- From the ground up, consolidation of power in consolidation, consolidation -> power -> profit
- DLedger on top of Named Data Networking (NDN): named-data.net since 2006.
- Annual NDN community meeting, 5-9 Sept 2019 Washington
Q: decentralising accountability, not in terms of security as access control. Q: Two different concepts: Proof of work/waste, identity (to enter).
- planning interim meeting – suggestions for venues wanted, please use mailing list
Followed by Side-Meeting on financial technology