Hi Jeff,
Regarding point 3. following thoughts:
- The owner, admin, or user of a thing has to trigger an update…their might be services that do the update on behalf
- In general we need an update mechanism, if e.g. an owner changes, it should be changed in discovery/search...not a big deal. Isn’ it?
From: dg-idot-bounces@kantarainitiative.org [mailto:dg-idot-bounces@kantarainitiative.org] On Behalf Of j stollman
Sent: Freitag, 24. Juli 2015 19:21
To: Alessandro Festa
Cc: dg-idot@kantarainitiative.org
Subject: Re: [DG-IDoT] Common identity standard
I am with Alessandro in the complexity of this solution in the real world.
- An iPhone is a collection of IoT devices (camera, audio recorder, touch screen, telephone, computer, etc.). Should each of these have its own "good key pair"? If not how do we handle the sale of just the camera by the same OEM who sells the camera to Apple? Do we need a way to aggregate devices?
- Separately, what constitutes a "good key pair"? Will all of the many unenlightened, non-high-tech manufacturers in the world participate? What is the likelihood that they will create duplicate key pairs when there are billions of devices? We tend to consider that we are servicing an environment where everyone is paying attention to international standards. Standards in markets as broad as we are discussion take decades to become pervasive. How many types of screws do we have? It isn't just metric versus "standard." Screws differ in diameter, pitch, head shape (flat, pan, etc.), and driver type (straight blade, phillips, head, star, etc.). And then there are custom screws. In IoT we will have hobbyist-types creating devices, along with old-line manufacturers. It isn't just an Apply and Samsung world.
- To Ingo's comment about relationships, how do we track changes in those relationships without creating a massive infrastructure? What happens when company A has a device that is used by employees A1, A2, and A3, sells the device to company B for use by B7, B8, and B9?
---------------------------------
Jeff Stollman
stollman.j@gmail.com
1 202.683.8699
Truth never triumphs — its opponents just die out.
Science advances one funeral at a time.
On Fri, Jul 24, 2015 at 7:34 AM, Alessandro Festa <afesta@alfweb.com> wrote:
related to the private key embeded by manufacturer I am wondering who would embed what in the case of a multi-manufacturer.
1) thing created by original manufacturer : embed a priv key
2) thing crafted/customized (oem) by second manufacturer : embed a priv key
when thing will need to act on behalf I expect to reflect a 1 to many relationship at this point and so I'll need as user to decide the degree of relationship between the various keys or only one single key pair will be allowed and this means we need to define a hierarchical policy to decide who will embed what.
I immagine an onion ring model based on user consent and relationship constrain: user to seller, seller to manufacturer (original or oem), manufacturer (oem) to manufacturer
Hi nat, I would follow on to your steps below
On 7/24/15 4:56 AM, Nat Sakimura wrote:
Yeah, it is nice, but WSDL would be too big.
Remember that sending 1 byte over the radio takes as much power as encrypting 1000 bytes. Also, memory and processing power is becoming cheap, so in IoT context, we should probably treat "minimizing the radio packet" as the priority.
As to the identification of the things are cocerned, the viable model that I imagine is as follows:
- The device manufacutrer creates a good keypair and embeds the private key (and its key thumbprint) in the device.
- For device authentication, use the key to sign the message.
When acting on behalf of a user
3. Authenticated user facilitates delivery of tokens to device
4. Device authenticates to AS using embedded keys in order to obtain tokens
5. Device uses tokens to authenticate to cloud endpoints, other device etc
Tokens thereby reflect 'relationship' of user & device
2015-07-22 1:33 GMT+09:00 Aninda Bhunia <abhunia@inc38.com>:
It would be interesting if we could create a standard that would allow even non IP devices to publish their identity through a wsdl type structure. Even if they are non IP at some point in their upwards relationship hierarchy their master gateway would be IP based and could be responsible for publishing the identity wsdls for the entities it brokers.
Thoughts ?
On Jul 21, 2015 11:52 AM, "Joni Brennan" <joni@kantarainitiative.org> wrote:
Noting I have no vote =)
I agree with Paul and others regarding discovery as the key initial mechanism. I believe Ingo has also noted this in the summaries from IDoT. Sal mentions NMAP / SNMP are there other exiting approaches? (apologies if this has been discussed in detail already)
Connecting Identity for a more trustworthy Internet - Overview
On Tue, Jul 21, 2015 at 8:42 AM, Salvatore D'Agostino <sal@idmachines.com> wrote:
Other than ip devices? In that case there are mechanisms support scanning ( eg NMAP) or SNMP that have been around for a while these are typically not exactly API friendly but do provide a starting point and we make good use in our offerings.
Salvatore D'Agostino
IDmachines LLC |1264 Beacon Street, #5
Brookline, MA. 02446 | USA
(one of) what is needed is a standardized mechanism for devices to present their identity (and those humans for which they are acting) to other things, cloud endpoints & applications
On 7/16/15 2:38 PM, Ranjan Jain (ranjain) wrote:
Hope everyone is doing well. Just wanted to bounce a question which I’m consistently getting asked around Identity, IoT perspective. Is there any industry standard in place or in works which can be used as a common standard across multiple identities. What I mean by this is that humans have SSN as an identity while a thermostat may have serial number while a network device may have a Mac ID as their identity. So, while individually they all have their own identity standard, when in the IoT world, all these entities start interacting with each other, how do we translate one identity into another or how will one identity interact with another identity in a standards way?
| | Cisco Systems, Inc.
400 East Tasman Drive
San Jose
California
95134
United States
Cisco.com | | |
Think before you print. | This email may contain confidential and privileged material for the sole use of the intended recipient. Any review, use, distribution or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender by reply email and delete all copies of this message. |
|
_______________________________________________
DG-IDoT mailing list
DG-IDoT@kantarainitiative.org
http://kantarainitiative.org/mailman/listinfo/dg-idot