This document summarizes a paper about naming in content-oriented architectures. It discusses two main naming types: self-certifying flat naming using cryptography, and user-friendly hierarchical naming. Key issues discussed are scalability, security, and flexibility. The document also covers how each naming type establishes security through integrity, confidentiality, provenance and availability. It compares the two approaches on usability, binding relationships, scalability through aggregation, and flexibility.
1 of 11
Download to read offline
More Related Content
Naming in content_oriented_architectures [repaired]
3. Scalability
Issues identified with Self-certifying names.
Not scalable as these are not hierarchical
Need a third party for translation purpose (like
DNS)
Claims:
Better scalability than hierarchical by using flexible
aggregation !
Have better security than human readable names.
Both naming architectures can be combined.
5. Entities for establishing security
Real world Identity (RWI) principal for data.
Name
Used for fetching the data
Provided by principal
Public key
Each principal associated with private/public key
Used for authentication purposes
6. Bindings
RWI Name: Helps in identifying the principal.
RWI Public key: Only intended principal could
claim to be the principal of data.
Public key Name: Associating key helps in
verifying the provenance of data.
Bindings are transitive proving any two of
them implies third one.
7. Bindings in Human readable, Hierarchical
names
Intrinsic binding:- RWI Name
Extrinsic binding:- Name - key by using certificate
authority (third party - PKI)
Bindings in self-certifying names (P:L)
Intrinsic binding (Name key):- Take hash of key at
any node and check whether it hashes to P.
Extrinsic binding:- RWI- key (external authority)
8. Comparison
Human readable, Hierarchical
(CCN)
More useful.
Remain unchanged as
cryptographic algos evolve.
Name - RWI binding not tight,
e.g., acronym ICSI.
Tiny URLs obliterate Name
RWI binding
Self-certifying, Flat (DONA)
Usability - Difficult to
remember such names
Require careful engineering to
retain names as algos. evolve.
9. Name Scalability
Hierarchical names help scalability
Reduce size of routing table
Reduce update rate of routing table
Flat Naming
Common Assumption Aggregation impossible
No hierarchy found but still supports greater
flexibility* in aggregation.
* Explicit aggregation is more flexible than inherent aggregation
10. Explicit aggregation
Deepest match working ??
Fetch Terms:
-Principal creates various fetch terms
corresponding to different third party
aggregates.
- Like outsourcing of content dissemination.
- Seems much like of CDN.
11. Naming and Flexibility
Human friendly naming require PKI for binding
keys to names.
Require universal agreement on the root trust
authority.
self certifying naming require external
authority for binding of keys to RWI.