mention graph isomorphism in security considerations

[pchampin]

@simoneonofri this is a followup on our conversation

---

Preview | Diff

[simoneonofri]

thank you @pchampin, this indicates potential Denial of Service issues (resource consumption). Might it also be worthwhile to specify the issue of Canonicalization on the Integrity side (e.g., in the case of the Verifiable Credentials)?

[pchampin]

thank you @pchampin, this indicates potential Denial of Service issues (resource consumption). Might it also be worthwhile to specify the issue of Canonicalization on the Integrity side (e.g., in the case of the Verifiable Credentials)?

This spec does not explicitly talk about canonicalization, which is the scope of RDF-CANON. But note that the new paragraph references the security considerations of RDF-CANON, where the mitigations are more developed (which is possible because RDF-CANON is about a specific algorithm, while the section in this PR has to remain very general, because that's the scope of RDF-CONCEPTS).