[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [openc2-lang] Comments of Note from the Actuator Profile subcommittee: New Language Element
DO we (open-c2) need to define spidx, swid, or other format? I think not, I think we incorporate be reference.
DO we (somewhere in OpenC2) have a list of the formats that are currently accepted? Likely yes.
DO they need to be discoverable, i.e., one can see a spidz, with a uri, to go and look up the format? I think not. Drilling down, if we start to have CBOR (say) Application Profiles, do we want them to reference the Language Spec for compatibility, or do we want them to reference the SBOM Application Profile for compatibility. I think the App Profile.
tc
From: openc2-lang@lists.oasis-open.org <openc2-lang@lists.oasis-open.org> on behalf of duncan sfractal.com <duncan@sfractal.com>
Sent: Thursday, December 12, 2019 12:28 PM To: Considine, Toby <Toby.Considine@unc.edu>; openc2-lang@lists.oasis-open.org <openc2-lang@lists.oasis-open.org> Subject: Re: [openc2-lang] Comments of Note from the Actuator Profile subcommittee: New Language Element
We have a substantive disagreement here wrt where I bolded your statement. I would like to specify, in the Language Specification, at least some of the standard formats. I would prefer every term in any OpenC2 Specification to be specified in the Language Specification and I would prefer, when possible, to even include terms in CAP’s when it would remove ambiguity and/or help with intertworking. So, for example, in the case of the SBoM formats; I would request swid, spdx, and cyclonedx be included in the language specification. I personally don’t care if swid, for example, is swid, swid_tag, software_id, or lots of other choices. But I do care that we pick one, and only one, term to represent it whenever it is used in any OpenC2 specification. And the place to do that definition is the language specification. I do agree CAP’s can extend the “list of formats” to longer than the 3 that I proposed. But even there, if someone went to the trouble to put a term in a CAP, then I would tend to favor adding that term to the language specification - unless they were truly obscure or blatantly vendor proprietary.
Duncan Sparrell sFractal Consulting LLC iPhone, iTypo, iApologize I welcome VSRE emails. Learn more at http://vsre.info/
From:
Toby Considine <Toby.Considine@unc.edu>
We are in substantial agreement. The language TC does not need to know if there are two, or three, or fifteen standard formats. That remains entirely up to the Application Profile.
As to the term report, it is a request to return a structured list, of more than one column, for what may well be more than a thousand similar elements, in a predefined format. I am not sure what the push-back is.
tc From: openc2-lang@lists.oasis-open.org <openc2-lang@lists.oasis-open.org> on behalf of duncan sfractal.com <duncan@sfractal.com>
Duncan Sparrell sFractal Consulting LLC iPhone, iTypo, iApologize I welcome VSRE emails. Learn more at http://vsre.info/
From:
openc2-lang <openc2-lang@lists.oasis-open.org> on behalf of Toby Considine <Toby.Considine@unc.edu>
As noted in the minutes of Monday’s meeting, the SBoM Actuator Profile is entering into rapid development. The Software Bill of Materials is a critical component of the Comply to Connect The SBoM draft will come to the language SC as a set of new requirements for remote query. This may be the precursor to other Profiles for remote situation awareness (“reporting”). More on that in another note. For this message, I want to focus on a new language element, one for something that is a BLOB as far as OpenC2 knows. SBoM already has a few standard formats, and none of them are JSON. In the case of SBoM, they are XML based, and there is some hint that the next generation of those formats will skip JSON and go to CBOR. We probably want to restrict these blobs to text-based, using the character-sets already legal in OpenC2. To the Language SC, those formats do not matter, although they do to the specific Actuator Profile. In this case, the well-known formats are SPDX and SWID. In another request they could be something else. I think we have a language element with a name similar to “Payload”. The language spec makes no assertions as to the content or format of a Payload. The notion of defining a custom serialization of a specific payload for OpenC2, one which must then be converted back into the original format for use was specifically rejected in the Actuator Profile call. A conforming Actuator Profile may request a response that has one or more Payloads. A Profile expecting a Payload MUST define the conformance expectations for the Payload. It is preferred that the conformance expectations be defined by reference to an existing specification. (This last point—pre-existing specification. Is it required? If there is no pre-existing specification, why would we define in a format other than JSON?) tc
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]