IPP Mail Archive: IPP> IPP Bake-Off 3 Summary

IPP> IPP Bake-Off 3 Summary

From: Zehler, Peter (Peter.Zehler@usa.xerox.com)
Date: Fri Oct 27 2000 - 14:02:06 EDT

  • Next message: Carl Kugler/Boulder/IBM: "Re: IPP> IPP Bake-Off 3 Implementation specific issues: BO3-IMP-4"

    All,

    I have attached the summary of the IPP Bake-Off 3. I have posted the entire
    document to the PWG site. The location is
    "ftp://ftp.pwg.org/pub/pwg/ipp/new_TES/Bake-Off-3-Summary.pdf". I seem to
    have problems just following the URL. I am able to retrieve the document
    with an FTP tool after being bounced several times due to number of
    connection limitations at the PWG site.

     The Complete results are being sent to the registered participants of
    Bake-Off. Any participant wishing to get the complete results should
    contact me.

    Pete
    ______________________________________________________________________
    The third IPP Bake-Off held October 17 to 20. It was hosted by Oak
    Technologies in Woburn Massachusetts. The Bake-Off was a success, though
    some participants wanted more time to test with the rich set of IPP
    implementations.
    Participating companies: Axent Technologies Inc., Canon, Electronics for
    Imaging Inc., Epson, IBM, i-data International, Japan Computer Industry,
    McAfee.com, Microsoft, Netreon Inc., NETsilicon Inc., Novell, Oak
    Technologies, Quality Logic, Ricoh, SEH Computertechnik Gmbh, Xerox
    The 18 participants provided 17 IPP Printers, 9 IPP Clients, 2 firewalls and
    2 HTTP Proxies. Out of the 153 possible combinations of Clients and
    Printers, 151 were tested.
            * The overall success rate was 93%.
            * Limiting the scope to IPP v1.1 provided a success rate of
    96%.
            * With IPP v1.0 Clients and v1.1 Printers, the success rate
    was 100%.
            * The tests with v1.1 Clients and v1.0 Printers resulted in a
    success rate of 31%, which is not surprising given that some printer
    implementations explicitly disallowed that combination. Some Clients were
    able to retry in v1.0 mode raising the success rate to 69%. It should also
    be noted that for v1.0 Printers that allow v1.1 Clients the success rate was
    100%.

    The majority of the failures can be attributed to one of two causes. The
    major cause of unresolved failures was due to IPP Clients that had problems
    with IPP Printers that sent HTTP "100 continue" messages. This was
    recognized as an implementation error. The other cause was v1.0 Printers
    that explicitly disallowed v1.1 Clients. IPP inherently provides a
    mechanism that allows minor version mismatches to be gracefully handled.
    The minor version mismatch was recognized as an unnecessary printer
    restriction.
    Security testing went well with both SSLv3 and TLS having no failures with
    their limited number of participants (8 and 7 respectively). Basic
    authentication had the most participants (59) with a success rate of 93%.
    The most common cause of failure here was the "100 continue" problem
    previously mentioned. Digest authentication was the poorest performer with
    31 participants and a success rate of 68%. Only a few of the failures were
    due to the "100 continue" problem.
    Firewall and HTTP proxy testing was a complete success. The testing with
    the firewalls demonstrated that administrators could set policy regarding
    IPP printing. Firewalls were able block, selectively allow or allow
    unrestricted printing between IPP Clients and Printers. The firewalls
    further demonstrated that they could add an additional layer of security
    requiring IPP Clients to authenticate to the firewall before allowing the
    IPP request through to a designated printer. The HTTP proxies operated
    transparently when used in IPP printing. No security interactions or
    caching issues were discovered.
    The IPP notification testing gave early implementers a chance to shake down
    their implementations. Out of the 25 combinations that were reported
    tested, only two complete failures were noted. The remaining 23 were able
    to subscribe for notifications. There were 19 successful "mailto"
    notification tests and 4 successful "INDP" tests. At least two issues with
    the notification documents were identified.
    The major benefit of any Bake-Off is bringing together the implementers of
    IPP from across the industry. The cooperation between the engineers was
    remarkable. All were sharing their IPP expertise and working together for
    the benefit of all. Every participating vendor will have an improved
    implementation of IPP as a direct result of this event.

    ______________________________________________________________________
                                    Peter Zehler
                                    XEROX
                                    Xerox Architecture Center
                                    Email: Peter.Zehler@usa.xerox.com
                                    Voice: (716) 265-8755
                                    FAX: (716) 265-8792
                                    US Mail: Peter Zehler
                                            Xerox Corp.
                                            800 Phillips Rd.
                                            M/S 139-05A
                                            Webster NY, 14580-9701



    This archive was generated by hypermail 2b29 : Fri Oct 27 2000 - 14:13:03 EDT