1. 1 IP Services over an ICN The Internet as our ICN use case Dirk Trossen, InterDigital Europe IRTF ICNRG Meeting © 2015 InterDigital, Inc. All Rights Reserved.

2. 2 A Hypothesis-Driven ICN Story Our starting point: The Internet is the killer app for ICN Make this work and you have a strong starting point for ICN Our hypothesis: IP over ICN has the potential to run IP services better than in standard IP networks Show that this is true and you have a strong reason for deploy ICN! Our approach: Design, build, trial and test Clear set of quantitative and qualitative statements regarding our hypothesis © 2015 InterDigital, Inc. All Rights Reserved.

3. 3 Expected Benefits Coincidental multicast Re-introduce multicast in the light of proliferating HTTP unicast Direct path mobility No more anchor points Fast surrogate support Evolution of CDNs but NO universal caching! Fast node/link resilience Delay-constrained insertion of surrogates Important for current 5G discussions Democratization of CDNs Content provisioning by smaller content providers (no need for large-scale CDNs) © 2015 InterDigital, Inc. All Rights Reserved.

4. Basic Idea Interpret IP-based communication as Named Object Exchange 4 © 2015 InterDigital, Inc. All Rights Reserved. Send HTTP GET to facebook.com Register as facebook.com Receive HTTP GET IP Network Subscribe to “facebook.com” Receive HTTP GET ICN Network Publish (HTTP GET) to “facebook.com”

5. 55 © 2015 InterDigital, Inc. All Rights Reserved. ICN NAP IP NAP ICN Border GW IP-only Sender UE IP (BGP) IP ICN F IP RVZ SDN Switch TM L2 ICN PR ICN RT ICN TP ICN NAP ICN F IP IP-only Receiver UE IP-only Sender & Receiver UE L2 ICN SR SDN Controller FN S1S1 S1S1 IP A Gateway-based System Architecture TM : topology manager RVZ: rendezvous FN : forwarding node SDN Switch ICN functions present in several ICN flavors

6. scope identifier for outside ICN communication scope identifier for inside ICN communication One Issue for (Standard) Agreement Namespace for HTTP over ICN mapping 6 © 2015 InterDigital, Inc. All Rights Reserved. IO root identifier FQDN:host.myname.tld URL:host.myname.tld/path/resource Hash(FQDN)Hash(URL) Hash(FQDN)Hash(URL) CID rCID

7. 7 From a Use Case to a Working, Deployable System Standardization Needs Various building blocks System architecture (possibly more the space for ETSI or others?) Namespaces for protocol abstractions Governing of root scopes (assignment? Scopes of scopes?) Interfaces/protocols Protocol encapsulations Lightweight transport (for HTTP abstraction) ICN-internal interfaces (particularly for highly delay-optimized versions) SB interfaces, particularly to SDN Main issue: IPoverICN proposition itself Use case that is valid to include in wider ICNRG discussionsOR Warranting an own work area? © 2015 InterDigital, Inc. All Rights Reserved.

8. 8 Summary ICN is an ongoing evolution We have moved from true endpoint communication to content oriented communication a long time ago We believe that the current Internet is the ultimate use case for ICN Make it work – great! Make it work better – potential for ICN deployment! We are dedicated to making IP/HTTP over ICN work KPI-driven approach (with trials) to quantify the improvement part © 2015 InterDigital, Inc. All Rights Reserved.

9. A Set of Partners (and Looking for More…) © 2015 InterDigital, Inc. All Rights Reserved. POINT (3.5MEuro H2020 effort) PartnerRole Academia (coordinator) Technology provider (technical mgr) Operator (trial lead) Vendor Content provider Ell.iIoT provider Academia (development lead) Academia (evaluation lead) Athens Univ. of Economics & BusinessAcademia (architecture lead) RIFE (3.0MEuro H2020 effort) PartnerRole SME (coordinator) Academia (technical mgr) Technology provider (architecture lead) Community operator (trial lead) Sat vendor Sat operator Academia (evaluation lead) Additional trial in Bristol (UK), …