I would suggest that the security issues of a Resource Service (and the
Resource Service type aspects that may be within any hardcopy device) are a
reasonable subject for the IDS group. At this point, my understanding is
that we are doing a conceptual modeling of MDF Services. As such, we must
recognize that there are security issues but it may not be reasonable to go
into them to a detailed but inevitably inadequate level.
Indeed, one could argue at what point do we grind P2600 considerations into
the models of all of the MFD services? If we decide it belongs in the
Overall MFD discussion (which may be justifiable), I withdraw from being
responsible for that document.
Security issues need to be address (along with several other things), but I
suggest that we establish the basic imaging model first.
Bill Wagner
From: mfd-bounces at pwg.org [mailto:mfd-bounces at pwg.org] On Behalf Of
nchen at okidata.com
Sent: Thursday, May 14, 2009 5:27 PM
To: Ira McDonald
Cc: mfd at pwg.org
Subject: [MFD] Re: Resource Service updates for Security
Hi Ira,
I am not sure why Health Assessment does not detect a server that has been
attacked and had its antivirus compromised. I certainly hope that the
situation would have been detected before MFD needed to scan virus.
What I want to say is that if a server has good security on guard, has
antivirus, and Health Assessment, and MFD has its P2600 standard level of
security on guard, then I don't think MFD's antivirus is that necessary.
-Nancy
Ira McDonald <blueroofmusic at gmail.com>
05/14/2009 04:46 PM
To
nchen at okidata.com, Ira McDonald <blueroofmusic at gmail.com>
cc
mfd at pwg.org
Subject
Re: Resource Service updates for Security
Hi Nancy,
The MFD does NOT have access to the current Health Assessment
for the upstream enterprise server - if that server has been attacked
and had its antivirus compromised, then it behooves the MFD NOT
to trust the firmware blindly - the essence of security is "trust nobody
more than you have to" - desktops (and now mobile devices) need
antivirus because sometimes the INFRASTRUCTURE systems get
corrupted too.
That was my point.
If an MFD doesn't have sufficient free-standing self-defense, then
sooner or later that MFD will be quarantined during Network
Endpoint Attachment and Health Assessment by a site policy that
doesn't accept the premise that "security is out-of-scope".
Cheers,
- Ira
Ira McDonald (Musician / Software Architect)
Chair - Linux Foundation Open Printing WG
Blue Roof Music/High North Inc
email: blueroofmusic at gmail.com
winter:
579 Park Place Saline, MI 48176
734-944-0094
summer:
PO Box 221 Grand Marais, MI 49839
906-494-2434
On Thu, May 14, 2009 at 4:31 PM, <nchen at okidata.com> wrote:
>> Hi Ira,
>> I hope the enterprise server itself already has antivirus software that
can
> detect the corruption before MFD retrieve the file. Otherwise, the
antivirus
> does not work well on the enterprise server. If that's the case, what's
the
> use to install antivirus software on MFD then?
>> -Nancy
>>>> *Ira McDonald <blueroofmusic at gmail.com>*
>> 05/14/2009 04:09 PM
> To
>nchen at okidata.com, Ira McDonald <blueroofmusic at gmail.com> cc
>mfd at pwg.org Subject
> Re: Resource Service updates for Security
>>>>> Hi Nancy,
>> Right - my point is, after you verify the digital signature,
> prudent security implementation is to SCAN the firmware
> before making it your next boot image - what if the correct
> digital signature comes from an enterprise server that has
> ITSELF become corrupted? Digital signature's no help.
>> Cheers,
> - Ira
>> Ira McDonald (Musician / Software Architect)
> Chair - Linux Foundation Open Printing WG
> Blue Roof Music/High North Inc
> email: blueroofmusic at gmail.com> winter:
> 579 Park Place Saline, MI 48176
> 734-944-0094
> summer:
> PO Box 221 Grand Marais, MI 49839
> 906-494-2434
>>> On Thu, May 14, 2009 at 3:00 PM, <nchen at okidata.com> wrote:
>> >
> > Hi Ira,
> >
> > Yes, trusted path, with encryption and signature whenever necessary. In
> > your case, you need to check the signature before executing.
> >
> > -Nancy
> >
> >
> >
> > *Ira McDonald <blueroofmusic at gmail.com>*
> >
> > 05/14/2009 02:55 PM
> > To
> > nchen at okidata.com, Ira McDonald <blueroofmusic at gmail.com> cc
> > mfd at pwg.org Subject
> > Re: Resource Service updates for Security
> >
> >
> >
> >
> > Hi Nancy,
> >
> > Trusted path for firmware merely means you know which
> > upstream server was corrupted - at least, if you CHECK
> > the content of the firmware before blindly executing it.
> >
> > And user scripts (and PostScript-like PDL jobs) can be
> > highly effective Denial-of-Service attacks (by crashing the
> > entire MFD or just the Print Service) unless scanned by
> > antivirus/antimalware tools.
> >
> > Those are real threats.
> >
> > Cheers,
> > - Ira
> >
> > Ira McDonald (Musician / Software Architect)
> > Chair - Linux Foundation Open Printing WG
> > Blue Roof Music/High North Inc
> > email: blueroofmusic at gmail.com> > winter:
> > 579 Park Place Saline, MI 48176
> > 734-944-0094
> > summer:
> > PO Box 221 Grand Marais, MI 49839
> > 906-494-2434
> >
> >
> > On Thu, May 14, 2009 at 2:42 PM, <nchen at okidata.com> wrote:
> >
> > >
> > > Hi Ira,
> > >
> > > P2600 covers the security threats that you mentioned here.
> > >
> > > In short, firmware update must be through trusted path.
> > >
> > > User's document if needed (depending on site security requirement),
> must
> > be
> > > protected from disclosure and/or alteration.
> > >
> > > There are many techniques to protect document disclosure and
> alteration.
> > We
> > > think the best techniques to prevent from disclosure is by encryption,
> to
> > > detect alteration is by signing digital signature.
> > >
> > > Many MFPs today do not have these security in place, but they will
> > > eventually when they comply to the security standard.
> > >
> > > -Nancy
> > >
> > >
> > >
> > > *Ira McDonald <blueroofmusic at gmail.com>*
> > >
> > > 05/14/2009 01:56 PM
> > > To
> > > nchen at okidata.com, Ira McDonald <blueroofmusic at gmail.com> cc
> > > mfd at pwg.org Subject
> > > Re: Resource Service updates for Security
> > >
> > >
> > >
> > >
> > > Hi Nancy,
> > >
> > > OK - I'm fascinated.
> > >
> > > What conceivable system security could replace the
> > > need for onboard antivirus software?
> > >
> > > Especially given that many MFDs now update firmware
> > > via a "special" Print Job - and do NOT have digitally
> > > signed firmware packages with an unbroken, network-
> > > accessible chain of authority to a major public Certificate
> > > Authority.
> > >
> > > Certainly scripts and macros invoked in ordinary
> > > Print Jobs are very easily corrupted - users create
> > > them, so they aren't digitally signed - verification of
> > > source does NOT remove the need for verification
> > > that the source system was not in fact corrupted.
> > >
> > > Cheers,
> > > - Ira
> > >
> > > Ira McDonald (Musician / Software Architect)
> > > Chair - Linux Foundation Open Printing WG
> > > Blue Roof Music/High North Inc
> > > email: blueroofmusic at gmail.com> > > winter:
> > > 579 Park Place Saline, MI 48176
> > > 734-944-0094
> > > summer:
> > > PO Box 221 Grand Marais, MI 49839
> > > 906-494-2434
> > >
> > >
> > > On Thu, May 14, 2009 at 10:54 AM, <nchen at okidata.com> wrote:
> > >
> > > >
> > > > Hi Ira,
> > > >
> > > > I haven't seen any security certification requirement that requires
> > > onboard
> > > > antivirus on MFP yet. Not sure what's the basis of your prediction.
> > > > Antivirus software vendors of course will try to push their products
> on
> > > any
> > > > platform. If you have not implemented MFP security standard,
properly
> > > > configure the system to secure it, antivirus might be a good idea to
> > > have.
> > > > But once you have proper security instrumented, antivirus is not
that
> > > > critical. It's nice to have of course.
> > > >
> > > > -Nancy
> > > >
> > > >
> > > >
> > > > *Ira McDonald <blueroofmusic at gmail.com>*
> > > >
> > > > 05/13/2009 06:20 PM
> > > > To
> > > > nchen at okidata.com, Ira McDonald <blueroofmusic at gmail.com> cc
> > > > mfd at pwg.org Subject
> > > > Re: Resource Service updates for Security
> > > >
> > > >
> > > >
> > > >
> > > > Hi Nancy,
> > > >
> > > > Several vendors currently offer MFDs that incorporate virus
> > > > scanning of all files and executable modules and that CAN
> > > > integrate into enterprise antivirus update schemes.
> > > >
> > > > Anyway, it was only an EXAMPLE (not a requirement) - though
> > > > I think you will find that network endpoint attachment protocols
> > > > and architectures are ALL going to require that you certify
> > > > that you have healthy onboard antivirus on your MFD in the
> > > > very near future.
> > > >
> > > > MFDs running Linux or other general-purpose operating systems
> > > > are just too dangerous anymore without antivirus (or firewalls).
> > > >
> > > > Cheers,
> > > > - Ira
> > > >
> > > > Ira McDonald (Musician / Software Architect)
> > > > Chair - Linux Foundation Open Printing WG
> > > > Blue Roof Music/High North Inc
> > > > email: blueroofmusic at gmail.com> > > > winter:
> > > > 579 Park Place Saline, MI 48176
> > > > 734-944-0094
> > > > summer:
> > > > PO Box 221 Grand Marais, MI 49839
> > > > 906-494-2434
> > > >
> > > >
> > > > On Wed, May 13, 2009 at 5:49 PM, <nchen at okidata.com> wrote:
> > > >
> > > > >
> > > > > Hi Ira,
> > > > >
> > > > > I agree that
> > > > >
> > > > > "The threat is EXECUTING those Resources on an Imaging
> > > > > System (the one hosting the Resource Service or ANY other
> > > > > Imaging System) or even a desktop client system."
> > > > >
> > > > > and
> > > > >
> > > > > "The Resource Service spec should NOT say anything,
> > > > > anywhere about internal versus external storage and
> > > > > should NEVER reference a Resource Repository."
> > > > >
> > > > > But I don't think "virus scan" is practical for an Imaging System
> > like
> > > an
> > > > > MFD or HCD to verify whether the stored or retrieved Executable
> > > Resources
> > > > is
> > > > > safe to execute. There are many other techniques to verify the
> > > integrity
> > > > of
> > > > > these type of resources that are more practical for an MFD or HCD.
> > > > >
> > > > > Did I misunderstand something?
> > > > >
> > > > > -Nancy
> > > > >
> > > > >
> > > > >
> > > > > *Ira McDonald <blueroofmusic at gmail.com>*
> > > > >
> > > > > 05/13/2009 05:01 PM
> > > > > To
> > > > > nchen at okidata.com, Ira McDonald <blueroofmusic at gmail.com>
> > > > > cc
> > > > > mfd at pwg.org Subject
> > > > > Re: Resource Service updates for Security
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > Hi Nancy,
> > > > >
> > > > > Internal or external storage of the Resouces has NOTHING to
> > > > > do with the security threat in 11.4 below.
> > > > >
> > > > > The threat is EXECUTING those Resources on an Imaging
> > > > > System (the one hosting the Resource Service or ANY other
> > > > > Imaging System) or even a desktop client system.
> > > > >
> > > > > The Resource Service spec should NOT say anything,
> > > > > anywhere about internal versus external storage and
> > > > > should NEVER reference a Resource Repository.
> > > > >
> > > > > Every reference in 11.1 through 11.3 to Resource
> > > > > Repository should be deleted. Users do NOT know
> > > > > where a Resource Service stores things - if a vendor
> > > > > does such an extension, then that vendor has broken
> > > > > good security practice in their system design and
> > > > > will justifiably fail independent security audits.
> > > > >
> > > > > The Resource Service is completely responsible *on its
> > > > > own* for every Resource that it stores (anywhere) or allows
> > > > > an authenticated user to retrieve (from the Resource
> > > > > Service, that is).
> > > > >
> > > > > Any network system that hosts a Resource Service
> > > > > is, by definition, an Imaging System - it may not be
> > > > > dedicated (single-purpose), but it's still an Imaging
> > > > > System.
> > > > >
> > > > > Another way of putting it is that "Imaging System"
> > > > > is NOT simply a synonym for "Multifunction Device"
> > > > > - it's a wider definition that includes Spoolers and
> > > > > network server-hosted Imaging Services.
> > > > >
> > > > > Cheers,
> > > > > - Ira
> > > > >
> > > > > Ira McDonald (Musician / Software Architect)
> > > > > Chair - Linux Foundation Open Printing WG
> > > > > Blue Roof Music/High North Inc
> > > > > email: blueroofmusic at gmail.com> > > > > winter:
> > > > > 579 Park Place Saline, MI 48176
> > > > > 734-944-0094
> > > > > summer:
> > > > > PO Box 221 Grand Marais, MI 49839
> > > > > 906-494-2434
> > > > >
> > > > >
> > > > > On Wed, May 13, 2009 at 12:16 PM, <nchen at okidata.com> wrote:
> > > > >
> > > > > >
> > > > > > Hi Ira,
> > > > > >
> > > > > > Thanks for the text for the Security Consideration.
> > > > > >
> > > > > > Here are something I need to clarify, and some comments.
> > > > > >
> > > > > > 11.4 - Security Threats from Executable Resources
> > > > > >
> > > > > > It's not clear to me whether the security problem described here
> is
> > > for
> > > > -
> > > > > > 1) storing Executable Resources at an external storage location
> > that
> > > > > > involves an external system for storing the resource, and the
> > > Resource
> > > > > > Service is hosted by an Imaging System. Or,
> > > > > > 2) storing Executable Resources internally within the Imaging
> > System
> > > > that
> > > > > > hosts the Resource Service.
> > > > > >
> > > > > > In both cases the security threat is to the Imaging System.
> > > > > >
> > > > > > If it's case 1) - I agree that the external system "SHOULD
verify
> > the
> > > > > > safety of such resources (e.g., by virus scanning)". But we
> agreed
> > in
> > > > > last
> > > > > > face-to-face meeting that it's out of scope of Resource Service
> to
> > > > > consider
> > > > > > anything related to the external storage system. What Resource
> > > Service
> > > > > > should consider is the restriction of the storage and retrieval
> > > > > operations
> > > > > > on such resources to authorized users. How the external storage
> > > system
> > > > > > should protect the stored executable resources is out of scope.
> > > > Although
> > > > > I
> > > > > > prefer to recommend some security objectives to be considered by
> > the
> > > > > > external storage system.
> > > > > >
> > > > > > If it's case 2) - I don't think we should give "virus scanning"
> as
> > an
> > > > > > example to an Imaging System to protect the stored Executable
> > > > Resources.
> > > > > > Virus scanning or intrusion detection techniques are common in
> PCs,
> > > but
> > > > > > rarely existent in Imaging Systems. It's simply not that
> practical
> > > for
> > > > an
> > > > > > Imaging System to have such software implemented which involves
> > > > constant
> > > > > > maintenance of upto-date virus signatures by external systems.
At
> > the
> > > > > > abstract level, we can provide some good security objectives to
> > > > consider
> > > > > for
> > > > > > the Imaging System, such as protection of confidentiality and
> > > integrity
> > > > > of
> > > > > > the resources and their related metadata. Like case 1), the
> storage
> > > and
> > > > > > retrieval of operations on such resources must to restricted to
> > > > > authorized
> > > > > > users.
> > > > > >
> > > > > > A Resource Service could be hosted by a computer remote to the
> > > Imaging
> > > > > > System that is under security consideration. We should consider
> > this
> > > > case
> > > > > > too.
> > > > > >
> > > > > > 11.5 - Security Threats from Static Resources
> > > > > >
> > > > > > Static resources that have associated Intellectual Property
> rights
> > or
> > > > > > license rights that involves metadata such as
> DateTimeOfExpiration
> > > > which
> > > > > > should also be protected for alteration. Therefore, not only
> > storage
> > > or
> > > > > > retrieval operations on Resources must be restricted, other
> > > operations
> > > > on
> > > > > > resource metadata must be restricted too.
> > > > > >
> > > > > > Aslo the security problem you described here is for storing
> Static
> > > > > > Resources internally in an Imaging System that hosts the
Resource
> > > > > Service.
> > > > > > We should also consider the case when the Resource Service could
> be
> > > > > hosted
> > > > > > by a computer remote to the Imaging System.
> > > > > >
> > > > > > Thanks,
> > > > > > -Nancy
> > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>----------------------------------------------------------------------------
-
> > > > > > Principal Engineer
> > > > > > Solutions and Technology
> > > > > > Oki Data
> > > > > > 2000 Bishops Gate Blvd.
> > > > > > Mt. Laurel, NJ 08054
> > > > > > Phone: (856) 222-7006
> > > > > > Email: nchen at okidata.com> > > > > >
> > > > > >
> > > > > >
> > > > > > *Ira McDonald <blueroofmusic at gmail.com>*
> > > > > >
> > > > > > 05/12/2009 03:15 PM
> > > > > > To
> > > > > > mfd at pwg.org, NancyChen <nchen at okidata.com>, Ira McDonald <
> > > > > > blueroofmusic at gmail.com> cc
> > > > > > Subject
> > > > > > Resource Service updates for Security
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > Hi Nancy, Tuesday (12
> May
> > > > 2009)
> > > > > >
> > > > > > Per my action from the Resource Service review during the April
> PWG
> > > > > > meeting, below is some text for the Security Considerations
> section
> > > of
> > > > > > the Resource Service.
> > > > > >
> > > > > > 11.4 Security Threats from Executable Resources
> > > > > >
> > > > > > Resources with a ResourceCategory of 'Executable' MUST be
handled
> > > with
> > > > > > special care by implementations of the Resource Service. Such
> > > > resources
> > > > > > can pose serious threats to the integrity of the Imaging System
> > that
> > > > > > hosts the Resource Service. In particular, such Resources can
be
> > > used
> > > > > > to introduce Trojan Horses to the Imaging System. If an
> > > implementation
> > > > > > of the Resource Service supports Executable resources, then that
> > > > > > implementation MUST restrict the storage of such resources
(e.g.,
> > to
> > > > > > authorized administrators and manufacturers) and SHOULD verify
> the
> > > > > > safety of such resources (e.g., by virus scanning).
> > > > > >
> > > > > > 11.5 Security Threats from Static Resources
> > > > > >
> > > > > > Resources with ResourceCategory of 'Static' SHOULD be treated
> with
> > > > > > special care by implementations of the Resource Service. Fonts
> > that
> > > > > > have associated Intellectual Property rights (e.g., as part of
> > their
> > > > > > network licenses) can pose serious threats to the availability
of
> > the
> > > > > > Imaging System that hosts the Resource Service - security audits
> > can
> > > > > > result in the shutdown or physical removal of the Imaging
System.
> > If
> > > > an
> > > > > > implementation of the Resource Service supports Static resources
> > that
> > > > > > have associated Intellectual Property rights, then that
> > > implementation
> > > > > > SHOULD restrict the storage of such resources (e.g., to
> authorized
> > > > > > administrators and manufacturers) and SHOULD restrict the
> retrieval
> > > of
> > > > > > such resources (e.g., to a configured group of authorized
users).
> > > > > >
> > > > > > Comments?
> > > > > >
> > > > > > Cheers,
> > > > > > - Ira
> > > > > >
> > > > > > Ira McDonald (Musician / Software Architect)
> > > > > > Chair - Linux Foundation Open Printing WG
> > > > > > Blue Roof Music/High North Inc
> > > > > > email: blueroofmusic at gmail.com> > > > > > winter:
> > > > > > 579 Park Place Saline, MI 48176
> > > > > > 734-944-0094
> > > > > > summer:
> > > > > > PO Box 221 Grand Marais, MI 49839
> > > > > > 906-494-2434
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > >
> > >
> > >
> > >
> >
> >
> >
>>>
--
This message has been scanned for viruses and
dangerous content by <http://www.mailscanner.info/> MailScanner, and is
believed to be clean.
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pwg.org/pipermail/mfd/attachments/20090514/c46f1e27/attachment-0001.html>