Re: [Roll] [6lowpan] New Version Notification for draft-qiu-roll-kemp: Do need an alternative security design ?

[Rene Struik <[e-mail]>]

Hi Qiu:

Perhaps, I was unclear: I did not suggest a "smart setting to tone down revocation". I did suggest that certificate revocation is only necessary in two cases, viz.
1) loss of secrecy and/or authenticity of the corresponding keying material.
2) change to authorization attributes associated with keying material. 
Case #1 occurs when one has a compromised device (extracted keying material via probing, side channel attacks, crappy security design, etc.). With Case #2, one should think of use cases typically quoted, such as employee receiving a public key issued by organization with authorization attributes and validity period of five years, and then discovering that employee left organization prior to certificate expiration set. (In my mind, not aligning lifetime of authorization with lifetime of cert is an example of bad implementation decisions by the organization [or plain laziness] and not necessary any more.)

With smart objects, I do not see those objects being reassigned continuously (a light switch usually assumes its role for its entire life), thus making reason #2 less likely as a trigger for change. Reason #1 could occur though, since objects can be expected low-cost where one cannot necessarily assume a trusted platform on board of the device (although, one could, e.g., wrap keys using a physical unclonable function and only unwrap during use). Handing out short-lived certs to devices, where assigning a new lease of life depends on some routines involving authentication and, e.g., inspection of whether the device's external casing has been damaged could then do the job. Short-lived certificates in a networked environments are perfectly feasible and per-device cost for internet of things environment should be almost zero (subsidies to support CA's lifestyle aside). 

One final note: once again, it is a misconception to only consider revocation issues with public keys. This topic area is independent of the type of key and applies to symmetric-key keying material as well. (In my mind, the security profession has not done such a terrific job here by mostly ignoring the topic in the symmetric-key case and coming up with umpteen schemes in the public-key setting, not much of them being practical.)

I hope this helps.

Best regards, Rene

==
In practice, the main
reason for revocation would usually be an authorization change and not
so much a key compromise setting. If so, revocation can be toned down
in smart object setting, since devices can be expected not to change
affiliation that often. On the other hand, devices are less well
protected, so key compromise may happen (if one does not implement key
security and implementation security with care).


On 11/2/2012 1:44 AM, QIU Ying wrote:
> Hi, Rene
>
> The idea of smart setting to tone down the revocation is very interesting.
> As we know, the vocation issue is a big challenge in security. It is
> appreciated that you could describe a bit more in detail.
>
> Regards and Thanks
> Qiu Ying
>  
>
>> -----Original Message-----
>> From: Rene Struik [mailto:[e-mail]]
>> Sent: Thursday, November 01, 2012 9:12 PM
>> To: QIU Ying
>> Cc: [e-mail]; [e-mail]
>> Subject: Re: [6lowpan] New Version Notification for draft-qiu-roll-kemp:
>> Do need an alternative security design ?
>>
>> Hi Qiu:
>>
>> Thanks for your note. One quick note re key revocation: key lifecycle
>> issues are independent of the "color" of the key (i.e., whether the key
>> is symmetric-key or public-key). Interestingly enough, usually this
>> issue is conveniently forgotten when symmetric keys come along, while
>> inflated when the word public key is mentioned. In practice, the main
>> reason for revocation would usually be an authorization change and not
>> so much a key compromise setting. If so, revocation can be toned down
>> in smart object setting, since devices can be expected not to change
>> affiliation that often. On the other hand, devices are less well
>> protected, so key compromise may happen (if one does not implement key
>> security and implementation security with care).
>>
>> I put a marker in my calendar to revisit your draft in detail.
>> Meanwhile, have a good discussion at the IETF meeting next week.
>>
>> Best regards, Rene
>>
>> On 10/31/2012 12:56 PM, QIU Ying wrote:
>>> Hi, Rene
>>>
>>> Thanks for your comments.
>>>
>>> The discussion of using public keys in MIP6 WG was much more than the
>>> description in RFC 4225, e.g. the lack of global PKI, the key
>>> revocation, etc. These issues also restricted to accept the public
>> key
>>> schemes in MIPv6 since a mobile device are always roaming and lost
>> easily.
>>> Regarding the scalability, according to my understanding, for example
>>> IKE, a pre-configured security policy (SP), which based on the home
>>> address of mobile devices, is needed before IKE exchange procedure.
>>> The pre-configuration is lack of scalability as the visiting mobile
>>> devices could be from any locations or any domains.
>>>
>>> The IKE scheme is only solve the issue of authentication between the
>>> mobile device and the correspondent node. It cannot ensure that a
>>> mobile device is reachable from other nodes.
>>>
>>> "resource utilization": did you mean the limited capability of mobile
>>> devices? I cannot remember if there are a lot of words on the
>>> capability in the MIPv6 specification. I thought it is not practice
>> to
>>> involve the revocation checking in a mobile device. Anyway, the
>>> capability issue is much more sensitive in LLNs than in mobile
>> networks.
>>> Your observation is correct that "get lots of message traffic to/from
>>> this third party and its local neighbours" because need more hops. In
>>> KEMP protocol, using the base station as the trust third party is
>> only
>>> in the bootstraps phase (or at a specified interval).  In the
>>> following update phases, the distribution mode should be employed. In
>>> the distribution mode, the previous neighbour router is role as the
>>> trust 3rd party to introduce the moving sensor to the next neighbour
>>> router. In this case, the total hops could reduce to 3. By the way,
>> in
>>> the public key scheme, the extra messages / communications are
>> required when the certifications need to update.
>>> I hope that the above explanation could be express the actual concept
>>> of the
>>> MIPv6 authors, not just on my own understanding ;)
>>>
>>> Regards
>>> Qiu Ying
>>>
>>>
>>>> -----Original Message-----
>>>> From: Rene Struik [mailto:[e-mail]]
>>>> Sent: Tuesday, October 30, 2012 2:27 AM
>>>> To: QIU Ying
>>>> Cc: [e-mail]; [e-mail]
>>>> Subject: Re: [6lowpan] New Version Notification for draft-qiu-roll-
>> kemp:
>>>> Do need an alternative security design ?
>>>>
>>>> Hi Qiu:
>>>>
>>>> Just curious: could you elaborate a little bit on the RFC 4225,
>>>> Section
>>>> 5.2 remark below? I just would like to understand scalability,
>>>> resource utilization, and other issues somewhat better and may have
>>>> missed something here. In particular, if one uses a symmetric-key
>>>> scheme with online involvement  of a trusted party who distributes
>>>> pairwise keying material, doesn't one then get lots of message
>>>> traffic to/from this third party and its local neighbors for each
>> protocol instantiation?
>>>> On a more general note, agreed there is a need to tackle trust life
>>>> cycle management in a dedicated forum. Originally, I thought the
>>>> Smart Object Security Workshop (which we had end of March 2012, just
>>>> prior to the IETF meeting) would be a good forum to tackle issues,
>>>> but felt we missed some opportunities there to bring forward an
>>>> agenda of things to accomplish (in my mind, there was too much
>> inside
>>>> the box thinking in terms of "tweaks to IETF drafts"), with less
>>>> emphasis on what makes ubiquitous networking different from a
>>>> deployment use case perspective (e.g., the lighting use case example
>> comes to mind).
>>>> Unfortunately, I will not be at the Atlanta meeting, though I might
>>>> be in Vancouver. Glad to contribute to call to action there.
>>>>
>>>> Best regards, Rene
>>>>
>>>> On 10/29/2012 12:03 PM, QIU Ying wrote:
>>>>> Dear all,
>>>>>
>>>>> Do need an alternative security design instead of the current
>> public
>>>> key protocols in key establishment? It's one of arguments in
>> previous
>>>> WG meeting.
>>>>> My answer is yes. Actually, the similar discussion had been raised
>>>>> in
>>>> mobile IPv6 WG (RFC4225).
>>>>> Besides the authentication, another major check is the reachability
>>>> checking to verify if the claimed mobile node is reachable (section
>>>> 4.1). RFC4225 also explains why the current Public Key
>> Infrastructure
>>>> (i.e. IKE) is not accepted in mobile IPv6 (section 5.2).
>>>>> Frankly, the scheme used in KEMP is not fresh new. It is in style
>> of
>>>> the popular Kerberos. Instead of sending the ticket to visiting
>>>> server from client directly in Kerberos, the ticket is sent to the
>>>> visiting server (new nearby router in KEMP) from the KDC (base
>> station in KEMP).
>>>> The benefit of this modification includes: 1) reduce the
>>>> communication;
>>>> 2) the client (mobile node in KEMP) is check if reachable from the
>>>> 3rd party (new nearby router); 3) revocation in time.
>>>>> Thank to many WG participants commenting on the draft (inclusive
>>>>> Rene
>>>> Struik, Steve Childress, Shoichi Sakane, Greg Zaverucha, Matthew
>>>> Campagna), the draft should be more mature and stronger.
>>>>> Regards
>>>>> Qiu Ying
>>>>>
>>>>>
>>>>>> -----Original Message-----
>>>>>> From: QIU Ying [mailto:[e-mail]]
>>>>>> Sent: Tuesday, October 23, 2012 11:57 AM
>>>>>> To: '[e-mail]'; '[e-mail]'
>>>>>> Subject: FW: New Version Notification for
>>>>>> draft-qiu-roll-kemp-02.txt
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> The KEMP draft is updated. The messages in the draft will be
>>>>>> carried in KMP format proposed by IEEE802.15.9 working group so
>>>>>> that the
>>>> KEMP
>>>>>> protocol is compatible with IEEE802.15.9 and could be deployed in
>>>>>> layer 2.
>>>>>>
>>>>>> Regards
>>>>>> Qiu Ying
>>>>>>
>>>>>>
>>>>>> -----Original Message-----
>>>>>>
>>>>>> A new version of I-D, draft-qiu-roll-kemp-02.txt has been
>>>>>> successfully submitted by Ying Qiu and posted to the IETF
>> repository.
>>>>>> Filename:	 draft-qiu-roll-kemp
>>>>>> Revision:	 02
>>>>>> Title:		 Lightweight Key Establishment and Management
>>>>>> Protocol in Dynamic Sensor Networks (KEMP)
>>>>>> Creation date:	 2012-10-22
>>>>>> WG ID:		 Individual Submission
>>>>>> Number of pages: 20
>>>>>> URL:             http://www.ietf.org/internet-drafts/draft-qiu-
>> roll-
>>>>>> kemp-02.txt
>>>>>> Status:          http://datatracker.ietf.org/doc/draft-qiu-roll-
>> kemp
>>>>>> Htmlized:        http://tools.ietf.org/html/draft-qiu-roll-kemp-02
>>>>>> Diff:            http://www.ietf.org/rfcdiff?url2=draft-qiu-roll-
>>>> kemp-
>>>>>> 02
>>>>>>
>>>>>> Abstract:
>>>>>>    When a sensor node roams within a very large and distributed
>>>>>> wireless
>>>>>>    sensor network, which consists of numerous sensor nodes, its
>>>> routing
>>>>>>    path and neighborhood keep changing.  In order to provide a
>> high
>>>>>>    level of security in this environment, the moving sensor node
>>>> needs
>>>>>>    to be authenticated to new neighboring nodes as well as to
>>>> establish
>>>>>>    a key for secure communication.  The document proposes an
>>>> efficient
>>>>>>    and scalable protocol to establish and update the secure key in
>> a
>>>>>>    dynamic wireless sensor network environment.  The protocol
>>>>>> guarantees
>>>>>>    that two sensor nodes share at least one key with probability 1
>>>>>>    (100%) with less memory and energy cost, while not causing
>>>>>>    considerable communication overhead.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> The IETF Secretariat
>>>>> Institute for Infocomm Research disclaimer:  "This email is
>>>> confidential and may be privileged. If you are not the intended
>>>> recipient, please delete it and notify us immediately. Please do not
>>>> copy or use it for any purpose, or disclose its contents to any
>> other
>>>> person. Thank you."
>>>>> _______________________________________________
>>>>> 6lowpan mailing list
>>>>> [e-mail]
>>>>> https://www.ietf.org/mailman/listinfo/6lowpan
>>>> --
>>>> email: [e-mail] | Skype: rstruik
>>>> cell: [phone] | US: [phone]
>>> Institute for Infocomm Research disclaimer:  "This email is
>> confidential and may be privileged. If you are not the intended
>> recipient, please delete it and notify us immediately. Please do not
>> copy or use it for any purpose, or disclose its contents to any other
>> person. Thank you."
>>
>>
>> --
>> email: [e-mail] | Skype: rstruik
>> cell: [phone] | US: [phone]
> Institute for Infocomm Research disclaimer:  "This email is confidential and may be privileged. If you are not the intended recipient, please delete it and notify us immediately. Please do not copy or use it for any purpose, or disclose its contents to any other person. Thank you."


-- 
email: [e-mail] | Skype: rstruik
cell: [phone] | US: [phone]