250 likes | 259 Views
This document provides a summary of the HIP-RG.meeting at IETF 61, including discussions on the status update, experiment reports, new types of locators, NAT and Firewall Traversal for HIP, and more.
E N D
HIP-RG meeting, IETF 61 November 12, 2004 Tom Henderson {thomas.r.henderson@boeing.com} Pekka Nikander {pekka.nikander@nomadiclab.com}
Agenda • 5 min Agenda bashing • 5 min RG status update • 20 min Report from HIP and Related Architectures workshop • 15 min HIP Experiment Report • draft-irtf-hip-experiment-00.txt • 10 min New types of locators • 15 min HIP Resolution and Rendezvous Problem Description • draft-eggert-hiprg-rr-prob-desc-00.txt • 15 min NAT and Firewall Traversal for HIP • draft-tschofenig-hiprg-hip-natfw-traversal-00.txt • 15 min Exchanging Host Identities in SIP • draft-tschofenig-hiprg-host-identities-00.txt • 30 min Open mike discussion on HIP deployment or other topics
RG status update • Mailing list: • http://honor.trusecure.com/mailman/listinfo/hipsec-rg • Supplemental web page: • http://hiprg.piuha.net/ • Charter • Evaluate benefit/impact of deploying HIP • Prepare report to IESG • Modus operandi: Gain experience on HIP deployment • Experiment with software • Analyze HIP in context of real networks
RG drafts Several have asked whether we can have RG documents? • We are chartered for at least one such document • HIP Experiment Report • We can add more based on RG consensus • Note: IRTF presently considering review and approval procedures for having a RG documents track
Logistics • Held on November 6 • By invitation only, based on white paper submittal and other considerations • 20 attendees • author from each submitted white paper • academic researchers from other projects (i3, NewArch, Delegation-Oriented Architecture, NIMROD) • RG chairs from Routing RG and DTN RG • IETF HIP, SIP, NSIS, MobIKE representation
Sessions • Applying and deploying an ID/locator split • changing and managing applications and hosts • dealing with legacy infrastructure and middleboxes • introducing new infrastructure • Overlays, rendezvous, middleboxes, and delegation • advanced middleboxes and firewalls • advanced resolution and indirection • General architectural directions • late binding • encouragement of middleboxes in architecture • approaches (FARA, HIP, i3, NIMROD, multi6, etc.)
Session 1: Deployment • Assume that users and networks want to deploy ID/locator separation • How to “cross the chasm” between architecture and reality (Early Adopters)? Deployed systems and workable infrastructure Architectures and specs
Session 1 organization • Host: Implementing and managing an ID/locator split • host and application concerns • Network: Making it work in today’s networks • firewalls • middleboxes (existing NATs) • (resolution) infrastructure • Incentives: Application/user incentives for deployment • what are the killer apps?
Session 1: Conclusions • Managing HIP is not a trivial task • HIP makes explicit some design choices that were implicit • We have probably not paid enough attention to middlebox traversal • a key deployment concern • No killer applications identified • Road Warrior and SIP explored • a framework for middlebox traversal? • or is HIP still a solution in search of a problem?
Session 2: advanced infrastructure • Maybe just one protocol (like in i3) • Maybe separated protocols (like HIP and ESP) • Maybe additional protocols • Registration, middle box internal, …
Session 2: Open questions • Rendezvous: overlay routing or name resolution? • Bootstrap: how to find an infrastructure node? • Layer 3.5 routing: • How much state in packet vs middle boxes? • How is the middle box state managed? • Effects of asymmetric routing? • What are the limiting and decisive factors?
Session 2: Open questions (2) • Address hiding and DDoS protection • Combination of different types of middle boxes? • Operations and management issues? • Debugging the system • Dangers of having any centralization • Aim for decentralised infrastructure? • How to manage free riding?
Session 3: Architecture • Discussion on other identifier types • Identity-Based Cryptography (Boneh-Franklin) • flat identifiers (i3) • Discussion of “what is HIP?” • A lot of functionality/features being overloaded into HIP • e.g. mobility management • Other architectural comments • late-binding of locators to identity
Summary of workshop • General value seen for HIP as “lowest location-independent identity in the stack” • are specific benefits enough to warrant deployment? • Recognize that HIP includes: A: public key to identifier binding (inherent) B: identifier binding to locators • Consider whether these can be separated, and different choices for A • Consider more carefully what is the “core HIP”
Summary of workshop (2) • How to coherently incorporate middle boxes? • Enumeration of what are the options • Discussion on legacy middle boxes and NATs • Killer apps? • NAT, FW, IPv4/v6 transition • or are there none? • Making configuration and management user-friendly is a hard problem
Experiment report draft-irtf-hip-experiment-00.txt • Intended to capture consensus for the IESG: • What are benefits and consequences of deploying HIP, or other ID/locator split? • What modifications to HIP are recommended? • Contributions needed from early adopters and implementors
Current skeleton outline 1. Scope 2. HIP Architectural Overview 3. HIP Architectural/Deployment Impact • on hosts (stack, APIs, management) • on infrastructure (DNS, firewall/NAT, advanced overlays) 4. HIP Experience • management, NAT traversal, scaling, deploying infrastructure, impact on applications, implementation experience, ...
Software status • Three current, public implementations of HIP available: • HIPL (HIP for Linux) (Helsinki HIIT) • FreeBSD HIP (Ericsson NomadicLab) • User-space Linux-based daemon (Boeing) • Max OS X and Windows XP under development • Boeing HIP testing server: • http://hipserver.mct.phantomworks.org • HIP infrastructure on PlanetLab (HIIT) • hi3 and OpenDHT integration
Software plans When can you start using HIP? • For bleeding edge types-- now • For early adopters (clean install, more user-friendly management)– maybe 6 months • we would like feedback on HIP usability and management (of current implementations)
Questions to RG • Do you favor continuing to meet on Friday afternoon of IETF? • How many people intend to experiment with HIP once software is more available?