[ Thanks to Kurt
Pfeifle for this report: ]
The Printer Working Group
(PWG) has finished the third Bake-Off to test the
interoperability of different implementations for IPP-enabled
products. The IPP (Internet Printing Protocol) has been designed by
the PWG and has the approval of the IETF (Internet Engineering Task Force)
in a set of RFCs (Request For Comments) which rate the IPP to be in
an “experimental” status for the time being.
The IPP, now in version 1.1 is supposed to become the successor
of the venerable LPR/LPD Protocol for network printing. LPR/LPD is
often thought of being based on RFC 1179. But in reality it never
was a formal standard. RFC-1179 itself describes its own contents
as “informational” only. In real life it nevertheless became a
de-facto standard, but was deviated from very soon by all its
implementers, almost beyond recognition and certainly beyond smooth
interoperability. As its shortcomings became felt more and more in
recent years, the vendors of printing hard- and software decided to
go for a new shot. And all the Big Birds of the industry joined in,
as you can see from a quick look into the PWG minutes.
IPP is designed to be based on and serve as an extension of HTTP
1.1 using its bi-directional abilities for GET-ting and
POST-ing data between a client and a server. This was no
bad decision at all from different points of view:
- HTTP 1.1 is a robust and proved protocol working in LAN and
WAN, through firewalls (or not), across continents and routers,
gateways; - vendor engineers have already quite a bit of experience with
HTTP as most of their network printers have an embedded webserver
inside already; - HTTP 1.1 will nicely be plugging into other internet standards
such as LDAP (for directory services) or SSL/TLS (for data
encryption).
The Common UNIX Printing System
(CUPS) which was selected by MandrakSoft to form the basis of a
new printing aera in their new 7.2 edition is exactly based on this
IPP. Also easyLinux 2.2 have adopted CUPS and many more vendors
will be stepping in soon…
At the Bake-Off there were 18 vendors present (Axent, Canon,
EfI, Epson, IBM, i-data, JCI, McAfee, Microsoft, Netreon,
NETsilicon, Novell, Oak, QualityLogic, Ricoh, SEH und Xerox), who
in between them had 17 IPP-printers, 9 IPP-clients and 2
HTTP-proxies. Out of possible 153 combinations for a pair of client
and printer there were 151 tested.
In the PWG’s Bake-Off
summary report they don’t tell you names of vendors or
products, they only give you tables. The overall success rate was
93 % for any client/printer combination. It was 100 % in the case
of 1.0 version clients and 1.1 version printers and a mere 31 % if
a 1.1 client was served by a 1.0 printer. This is not the least
dissapointing as some printer implementations explicitly dissallow
that combination.
It seems there were quite a few new products present at this
test field whereas on the other hand there were a few “proved ones”
missing: names as AXxis, Lexmark, HP, Minolta or Easy Software Products (CUPS) are
missing on the list of participants. Those vendors are named on the
PWG home page to contribute to the number of around 150
IPP-products available so far.
A very positive balance sheet is drawn from the first-ever
testings of the IPP-security features:
- there was not a single failure with SSL3 (Secure
Socket Layer) and TLS (Transport Layer Security); - Basic Authentification worked 93% (here an
implementation error on the part of different vendors was
recognised, caused by a “HTTP 100 continue” message); - Digest Authetification rated weakest with only 68%
success; - Firewalls did work as expected: block or let traffic pass
according to configuration, and also they processed the clients
authentifications with no problems; - HTTP proxies – switched in between an IPP client and an IPP
printer – worked without a flaw even if security was enabled
Now, do you have to be cautious if you come across that
brand-new “IPP inside” sticker when you unpack your next
network printer>? No, most IPP-enabled printers will still be
supporting the old-fashioned LPR/LPD thingy for years to come. But
I hope, Grant Taylor’s (Linux Printing HOWTO) new-born
Crusade for the Death
of LPD will succeed way before that…
And Mandrake 7.2 and easyLinux 2.2 gim-me hope!