EKG Identifier (EKG/ID)¶
Usually the EKG/ID is part of an EKG/IRI which makes it "web-resolvable" as well, but at higher levels of technical EKG maturity there may be other technical ways to ensure "resolvability". It may not be required anymore to mandate the use of the HTTP protocol for that, there may be other more advanced protocols that deliver full resolvability with additional features that cannot be provided by HTTP.
Not to be confused with Canonical Identifiers.
A universally unique opaque (i.e. meaningless) number, either based on a hash or a random number, optionally signed that uniquely identifies an object1 in the EKG.
An identifier is "opaque" if it provides no information about the thing it identifies other than being a seemingly random string or number.
An EKG/ID needs to be opaque because one of the most impactful and important objectives of EKG is to maximise proliferation of web-resolvable EKG Identifiers (i.e. EKG/IRIs) across not only the internal enterprise but also its larger ecosystem or even beyond that. These identifiers end up for years, perhaps decades, in all kinds of databases, spreadsheets, documents, long term storage etc. Therefore, no user or system should be able to derive any meaning or actual information from these identifiers.
The EKG/IRI identifier is permanent and can safely be proliferated across the enterprise's universe---including its ecosystem---and will be used for the expression of facts about the object including relationships between objects.
An EKG Identifier cannot be reused for another object. Once attached to a given thing in the EKG, it remains the identifier of that thing forever, even until after the end of the logical life-cycle of that thing.
Resolving an identifier can be done in three ways:
- using it in a transaction---i.e. a query or update statement---submitted or routed via an internet protocol (e.q. HTTP API) or other means (e.g. Kafka) to a "lookup service" that translates one or more given "features" of an object to its corresponding EKG Identifier (i.e. an EKG/ID or an EKG/IRI).
- constructing it via a standardized policy from key components and applying a hash and optionally signing it---where the object represented by the EKG identifier may or may not already exist.
- constructing it by giving the object an EKG identifier based on a random number in case the EKG is the authoritative source for the given object.
An EKG Identifier (EKG/ID) (currently, see side box) becomes "web-resolvable", by making it part of an EKG/IRI, i.e. by making it a URL and thereby prefixing the "identifier part" with "a locator":
[locator] could be something like
If an external dataset is loaded into your organization's EKG then you could decide not to give each object in that dataset to have a standard EKG/IRI and just use the provided identifiers, whatever they may be. However, that could mean that it will be technically more difficult to make these external IRIs resolvable to your organization's EKG and that the proliferation of those external identifiers across your data and technical landscape could create a dependency towards the organization that provided the dataset.