Jeffrey,
In TAXII land you also, more than likely be able to just ask for an Indicator or other TLO, without the overhead of the package. So something like:
{ "type": "indicator", ... }
Thanks,
Bret Bret Jordan CISSPDirector of Security Architecture and Standards | Office of the CTO Blue Coat Systems PGP Fingerprint: 63B4 FC53 680A 6B7D 1447 F2C0 74F8 ACAE 7415 0050 "Without cryptography vihv vivc ce xhrnrw, however, the only thing that can not be unscrambled is an egg."
Take a look here for the working definition of a package: https://docs.google.com/document/d/1YcEtyUGdFkJIPdDZ7K-mHbvjFt-5pOL2EIw_ZJuqNpM/edit#heading=h.c9oxowopqs2As you can probably see, it’s essentially your former suggestion. We’ve had discussions about it though and I’ve argued the latter, mainly for ease of use. Bottom line is that we haven’t really decided, and have gone with the 1.x approach for the time being. Also, Paul Patrick from iSight has provided this notional example: http://taxii2-demo.soltra.com/taxii/mygroup/collections/mycollection/packages/package--3b3441de-8bf2-409e-a7e8-8f296f385057In terms of validation…because of our `type` keyword it’s actually pretty easy to validate. The challenge is on understanding the validation message, because what you’ll get back is: you didn’t provide an attack pattern OR a malware OR an indicator OR a threat actor, etc. and it’s up to you to figure out which you actually wanted. John On 3/21/16, 2:55 PM, "Mates, Jeffrey CIV DC3/DCCI" < cti@lists.oasis-open.org on behalf of Jeffrey.Mates@dc3.mil> wrote: I have been trying to make sure I'm up to date on what a STIX 2.0 document will look like, and while there is a great deal of information about particular object types and common attributes I haven't had much luck finding an example of what the shell of a document will look like. Does anyone know if we have a generally agreed upon sample of this somewhere?
So far I have heard two different visions of STIX 2.0 the first more aligns to STIX 1.X and roughly maps to a json format of: { Header: [], Observables: [], Indicators: [], ... Relationships: [] }
The second moves to a node link model along the lines of:
{ Header: [], Objects: [], Relationships: [] }
I think that the second model makes lookups simpler when resolving relationships while also making adding new object types easier, but also may introduce additional challenges when attempting to validate the JSON's schema.
I haven't found confirmation on what has been generally agreed upon or if a consensus has been reached.
Jeffrey Mates, Civ DC3/DCCI ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Computer Scientist Defense Cyber Crime Institute jeffrey.mates@dc3.mil 410-694-4335
|