PS> Thoughts on Discovery for PSI

PS> Thoughts on Discovery for PSI

McDonald, Ira imcdonald at sharplabs.com
Wed Jun 12 14:48:27 EDT 2002


Hi folks,

It seems to me that PSI usage scenarios that require discovery (essentially
all of them) separate into two broad classes, just like discovery for IPP:

1)  Discovery of available 'public' print services (for example, while
staying
    in a hotel)
    - public registries of services like UDDI (the future practice, not the
      current state of deployment) would work well here
    - but UDDI doesn't give any help solving the 'locality of reference' 
      problem (that is, you don't want to find a 'public' print service
      in Timbuktoo - you want one _very_ close to your roaming handheld
      geographically)
    - SLPv2 (RFC 2608) with Gateway Directory Agents (an extension currently

      under development at Columbia University) would also work (if widely
      deployed in the future)

2)  Discovery of available 'private' print services (within a single 
    administrative domain - for example, an enterprise LAN)
    - SLPv2 (RFC 2608) would work well (using the existing 'printer'
      SLP schema)
    - LDAPv3 would work well (using the existing 'printer' LDAP schema,
      which is technically aligned with the SLP 'printer' schema)
    - DNS SRV (service) records would also work well (but they burden
      the enterprise DNS and administrators much more than SLPv2 does)
    - dozens of proprietary discovery services, like Microsoft SSDP,
      would also probably work (but they're usually impossible to 
      successfully deploy across a heterogenous enterprise LAN).

For those interested learning more about remote discovery via SLPv2 see:

ftp://ftp.ietf.org/internet-drafts/draft-zhao-slp-remote-da-discovery-03.txt
(26 March 2002)

Cheers,
- Ira McDonald
  High North Inc



More information about the Ps mailing list