Re: [dtn-interest] Delay tolerant networking and absolute timing

["Ivancic, William D. (GRC-RHN0)" <[e-mail]>]

Like most things, the answer is yes and no.

For example, if you are using something like BSP PIB with certificates, you need to be in sync enough to have the certificate not be expired or to be valid (not to early, not to late).  But, that is somewhat independent of the core Bundle Protocol (RFC5050).  

If you know you can never sync yours systems, you need to design your security mechanisms to handle that.  

--Will

>-----Original Message-----
>From: [e-mail] [mailto:dtn-
>[e-mail]] On Behalf Of Burleigh,
>Scott C (313B)
>Sent: Wednesday, September 30, 2009 4:07 PM
>To: [e-mail]
>Subject: Re: [dtn-interest] Delay tolerant networking and absolute
>timing
>
>Sorry, reviving this old thread because some debugging we were doing
>today raised a question I hadn't thought of.  For environments in which
>we can't synchronize clocks among nodes, for one reason or another, are
>there additional security issues?  I'm thinking of defense against
>replay attacks, coordination of the use of life-limited keys, that sort
>of thing.  Maybe that's only a concern for Kerberos and elsewhere has
>been superseded by other techniques?  Do those techniques fit well into
>a delay-tolerant architecture?
>
>Scott
>
>> -----Original Message-----
>> From: [e-mail] [mailto:dtn-
>> [e-mail]] On Behalf Of Henning
>Rogge
>> Sent: Thursday, August 06, 2009 7:07 AM
>> To: [e-mail]
>> Subject: [dtn-interest] Delay tolerant networking and absolute timing
>>
>> Hello,
>>
>> I have visited the workinggroup meeting in Stockholm and would like to
>> start a discussion about absolute timestamps in DTN (the last
>presentation
>> mentioned the possibility to adapt DTN to relative timestamps).
>>
>> Absolute timing can be a real pain in large networks with autonomous
>nodes.
>> Many embedded devices lack a GPS receiver and manets/mesh networks
>cannot
>> rely on a central time source like GSM or UMTS networks. Setting the
>time
>> to a "roughly synchronized" value doesn't really scale well with
>network
>> size. I just try to imagine an emergency response team after a tsunami
>> which has so sync 500 nodes before they can start, especially if they
>are
>> coming from 10 different countries distributed over a few days.
>>
>> I think it would be a good idea to switch to relative timestamps where
>it
>> is possible. If some parts of the protocol cannot work without
>absolute
>> timing, each station could try to learn the time difference to all of
>it's
>> neighbors (without sync the clocks), so it can translate incoming
>absolute
>> timings into local time.
>>
>> What do you think about this ?
>>
>> Henning Rogge
>>
>> *************************************************
>> Diplom-Informatiker Henning Rogge
>> Forschungsgesellschaft für
>> Angewandte Naturwissenschaften e. V. (FGAN) Neuenahrer Str. 20, 53343
>> Wachtberg, Germany
>> Tel.: [phone]
>> Fax: [phone]
>> E-Mail: [e-mail]
>> Web: www.fgan.de
>> ************************************************
>> Sitz der Gesellschaft: Bonn
>> Registergericht: Amtsgericht Bonn VR 2530
>> Vorstand: Dr. rer. nat. Ralf Dornhaus (Vors.), Prof. Dr. Joachim Ender
>> (Stellv.)
>
>
>_______________________________________________
>dtn-interest mailing list
>[e-mail]
>http://maillists.intel-research.net/mailman/listinfo/dtn-interest