ADR 003: DID Resolver | Product Docs (2024)




Alex Tweeddale, Renata Toktar, Ankur Banerjee, Andrew Nikitin

ADR Stage


Implementation Status

In Progress

Start Date


Last Updated



The did:cheqd method ADR defines how DIDs are created and read from ledger. According to the W3C DID Core specification, DID methods are expected to provide standards-compliant methods of DID and DID Document ("DIDDoc") production.

The cheqd DID Resolver is designed to implement the W3C DID Resolution specification for did:cheqd method.


The DID Resolution specification prescribes a defined algorithm with standardised behaviour for expected and unexpected inputs that a conforming DID method must be able to produce.

All conforming DID resolvers implement resolve and resolveRepresentation abstract functions, as defined in the DID Resolution specification.

Resolve function

The resolve function is intended to fetch the abstract form of the DID Document, as stored on the ledger. This abstract/raw form may not necessarily be in JSON/JSON-LD format as the underlying data persistence layer where the DIDDoc is stored for any particular method might use different serialisation/storage formats.

resolve(did, resolutionOptions) →« didResolutionMetadata, didDocument, didDocumentMetadata »

Since cheqd uses the Cosmos SDK blockchain framework, the underlying data storage and retrieval ("resolve") mechanisms used rely on those offered by the Cosmos SDK framework. Cosmos SDK uses Protobuf (Protocol Buffers) encoding for its wire protocol.

Cosmos SDK framework typically provides gRPC/gRPC-Web, JSON-RPC, and REST API endpoints for on-ledger modules and functionality.

For example, did:cheqd:testnet:DAzMQo4MDMxCjgwM can be fetched using the native Cosmos SDK REST API endpoint (or equivalent endpoints). This provides responses that would meet the abstract definition of a resolve function as defined in the DID Core specification.

In case of the cheqd network testnet, an instance of this resolve endpoint through the Cosmos SDK REST API would be which returns the following response:

{ "value": { "did_doc": { "context": [ ], "id": "did:cheqd:testnet:97e351e6-2d9d-4314-82ec-e0d12bc5de43", "controller": [ "did:cheqd:testnet:97e351e6-2d9d-4314-82ec-e0d12bc5de43" ], "verification_method": [ { "id": "did:cheqd:testnet:97e351e6-2d9d-4314-82ec-e0d12bc5de43#key-1", "verification_method_type": "JsonWebKey2020", "controller": "did:cheqd:testnet:97e351e6-2d9d-4314-82ec-e0d12bc5de43", "verification_material": "{\"crv\":\"Ed25519\",\"kty\":\"OKP\",\"x\":\"q8-CHj4_nIYo8tK5RdjYbXlsTUnwW_i4gIEclps2i2o\"}" } ], "authentication": [ "did:cheqd:testnet:97e351e6-2d9d-4314-82ec-e0d12bc5de43#key-1" ], "assertion_method": [ ], "capability_invocation": [ ], "capability_delegation": [ ], "key_agreement": [ ], "service": [ ], "also_known_as": [ ] }, "metadata": { "created": "2023-03-01T08:47:07.919899771Z", "updated": "2023-03-01T08:52:27.785774183Z", "deactivated": false, "version_id": "cfe2f51f-8ec5-4fd8-8ab9-61859de879f4", "next_version_id": "", "previous_version_id": "482feb56-1de8-4c36-bc00-a15a4188bccc" } }}

As you can see in the response body above, this is the raw Protobuf fetched from the cheqd testnet ledger, marshalled into a JSON form. Crucially, this form has certain deviations from the JSON/JSON-LD production expected in DID Core specification:

  1. JSON key names that correlate to DID Core properties are listed in snake_case, rather than camelCase as required. This is because Protobuf standard linting rules require these properties to be defined in snake_case.

  2. DID Core properties with empty values are still shown in this JSON, whereas the requirement is to drop them from standards-compliant DIDDoc representations.

Resolve Representation function

The resolveRepresentation abstract function, as defined in DID Core specification, is intended to address concerns similar to the ones highlighted above to product a standards-compliant JSON/JSON-LD representation of a DIDDoc.

resolveRepresentation(did, resolutionOptions) →« didResolutionMetadata, didDocumentStream, didDocumentMetadata »

For example, a resolveRepresentation function could derive a valid standards-compliant representation of did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8 from the above resolve function. The response would be similar to the one below containing Resolution Metadata, DIDDoc, and DIDDoc Metadata:

{ "didResolutionMetadata": { "contentType": "application/did+ld+json", "retrieved": "2022-08-03T09:52:49Z", "did": { "didString": "did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8", "methodSpecificId": "ea2b76cf-a118-403a-8f49-244e56c9dcb8", "method": "cheqd" } }, "didDocument": { "id": "did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8", "verificationMethod": [ { "id": "did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8#key1", "type": "Ed25519VerificationKey2020", "controller": "did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8", "publicKeyMultibase": "z6jVkB274neVf7iJETpMECwznBF8wDe8tpvF4BZLRZgMU" } ], "authentication": ["did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8#key1"] }, "didDocumentMetadata": { "created": "2022-07-19T08:29:07Z", "versionId": "a0a4f3a3-829d-4188-bc78-8bea6a522037", "linkedResourceMetadata": [ { "resourceURI": "did:cheqd:testnet:ea2b76cf-a118-403a-8f49-244e56c9dcb8/resources/44547089-170b-4f5a-bcbc-06e46e0089e4", "resourceCollectionId": "ea2b76cf-a118-403a-8f49-244e56c9dcb8", "resourceId": "44547089-170b-4f5a-bcbc-06e46e0089e4", "resourceName": "DemoResource", "resourceType": "CL-Schema", "mediaType": "application/json", "created": "2022-07-19T08:40:00Z", "checksum": "7b2022636f6e74656e74223a202274657374206461746122207d0ae3b0c44298", "previousVersionId": "", // empty string if no previous version, otherwise, resourceId of previous version "nextVersionId": "" // empty string if no new version, otherwise, resourceId of new version } ] }}

Architecture of DID Resolver for cheqd

As described above, the abstract resolve function is already available for the cheqd ledger via the default Cosmos SDK gRPC/REST API endpoints. Our primary objective with building a DID Resolver for cheqd was to design this resolveRepresentation piece as a standalone component that was not packaged within the cheqd-node ledger code.

This objective has certain advantages:

  • Updates to DID Resolver code can be carried out and released independently of cheqd-node releases. As a consequence, there's no need to go through an on-ledger governance vote, and voting period to make a change to resolveRepresentation.

  • A separate web service module would allow for flexibility in how to handle complex scenarios on DID URL Dereferencing, error code handling for DID URL requests, and safely handling content transport for various media types.

  • Making the DID Resolver a standalone, non-ledger module allows for an operator of this web service to independently scale their service horizontally and vertically.

We explored two architectural patterns for how a DID Resolver could be implemented for the cheqd ledger. The objective here was to explore and provide DID resolution operators multiple approaches for running resolution service, each with their own pros and cons (which are discussed below).

  1. "Full" cheqd DID Resolver

    1. Since the cheqd-node ledger / Cosmos SDK is written in Golang, this resolver would consist of Golang libraries imported from the existing ledger code. This promotes code reuse.

    2. Data would be fetched from the ledger using the gRPC endpoint on a node, which allows it (by default) to take place over an encrypted channel since gRPC uses HTTP/2.

    3. Data retrieved would be in the native Protobuf representation as stored on ledger, thus allowing data integrity computations to be made.

  2. "Light" cheqd DID Resolver

    1. Universal Resolver drivers are designed to be run as Docker containers. A limitation of this approach is that the computation footprint of a compute resource can be quite high, e.g., a Docker container may be 100 MB+ in size and suffer from slow startup times in a "cold-start" scenario.

    2. Thus, our "Light" DID Resolver idea was to explore using Cloudflare Workers, a lightweight serverless compute platform. As a comparison, Cloudflare Workers are limited to 1 MB in size and have extremely low cold-start times. (We use Cloudflare Workers in our Cosmos SDK Custom Data API, for example.)

    3. Cloudflare Workers can also be deployed outside the Cloudflare service in a Docker container using Miniflare. This could be used to provide a Docker container deployment option for the Universal Resolver did:cheqd driver.

    4. However, a limitation of Cloudflare Workers is they do not allow a gRPC request to be made to an external endpoint. This would force the "Light" cheqd Resolver to use the gRPC-Web / REST endpoint resolve implementation to fetch data from the ledger. This could be considered a higher risk profile in terms of data integrity by resolver operators / client applications.

Both of the architectural patterns above are designed so that a Universal Resolver driver for did:cheqd could be created. The Universal Resolver project aims to provide a common REST API definition for DID Resolution where each DID method can provide a Docker container that with an easy-to-deploy mechanism for the specific DID method.

Full cheqd DID Resolver

The Full cheqd DID Resolver is able to use as a Golang module for send resolve requests to a cheqd node instance to fetch DIDDoc / Resources from the ledger.

Since the Full cheqd DID Resolver is wrapped for usage as a Docker container image using the Universal Resolver specification, the end-to-end sequence diagram for our DID Resolver would look like below:

Figure 1: "Full" cheqd DID Resolver sequence diagram (editable version)

The Full cheqd DID Resolver is designed to handle requests concurrently, while reducing the risk of large quantities of threads and requests blocking the efficiency of the on-ledger services.

Fetching Protobuf from ledger and converting it to JSON

Since Cosmos SDK SDK encodes data in Protobuf, the DID Resolver "marshalls" them to JSON. The software class diagram below describes how these components/methods are tied together:

Figure 2: "Full" cheqd DID Resolver class diagram

Marshalling/unmarshalling requests back-and-forth between Protobuf and JSON is carried out by services in the "Full" DID Resolver

Figure 3: "Full" cheqd DID Resolver Protobuf <-> JSON marshalling (editable version)


Given the drawbacks associated with a Light cheqd DID Resolver being unable to send gRPC requests to a cheqd node instance, the decision was taken to (initially) implement the Full cheqd DID Resolver architecture. Future work might separately consider and design a Light DID Resolver profile that can work with Cloudflare Workers while also allowing deployment through Docker.

Universal Resolver driver for did:cheqd

Compiled packages/binaries for the Full cheqd DID Resolver will be made available as a Docker container image with standardised Docker Compose configuration as defined in the Universal Resolver driver development guide to provide an easy mechanism for DID Resolution operators to incorporate did:cheqd as one of the supported DID methods they handle.

Configuration properties for Universal Resolver driver

  1. Operators should be able to configure gRPC endpoints for mainnet as as testnet (or any other cheqd network namespaces) so that a single DID Resolver instance can resolve requests to any network. (E.g., grpc.cheqd.netfor mainnet, for testnet)

  2. For each supported cheqd network namespace (e.g., mainnet, testnet), operators should be able to define one or more upstream cheqd node gRPC endpoints. (E.g., additional fallback endpoints besides for mainnet.) In case any one of the upstream endpoints is unavailable, this allows the DID Resolver to try a different upstream gRPC endpoint for resiliency.

  3. Operators should be able to define whether their gRPC pull/request is secure (default, carried out over HTTP.2 with TLS handshake) or insecure (optional, equivalent to a grpcurl -plaintext connection downgrade). This accommodates scenarios where gRPC endpoints may not necessarily be equipped to handle TLS handshakes, since the default Cosmos SDK / cheqd node configuration endpoints do not have TLS certificates defined.


Backwards Compatibility

  1. Not applicable, since this would be the first release of cheqd DID Resolver


  1. Full cheqd DID Resolver reuses existing code to handle some parts of resolve and representing DIDDocs.

  2. Third-party applications that want to implement their own DID Resolver implementations for cheqd can consume the current implementation as a Golang module (e.g., import "").

  3. Universal Resolver driver implementation provides a standardised way of incorporating did:cheqd into Universal Resolver instances that support multiple DID methods.


  1. Use of a custom DID URL path (e.g., /1.0/identifiers/{did}/resources/{resource-id} can be non-standard). Efforts should be made to see if this can implemented instead using DID URL queries according to the Trust over IP Foundation "DID URL Resource Parameter" specification.


  1. Lack of a Light cheqd DID Resolver, at least initially, might not give an easy and computationally-cheap alternative to running Docker containers.


ADR 003: DID Resolver | Product Docs (2024)


Top Articles
Latest Posts
Article information

Author: Foster Heidenreich CPA

Last Updated:

Views: 5901

Rating: 4.6 / 5 (56 voted)

Reviews: 95% of readers found this page helpful

Author information

Name: Foster Heidenreich CPA

Birthday: 1995-01-14

Address: 55021 Usha Garden, North Larisa, DE 19209

Phone: +6812240846623

Job: Corporate Healthcare Strategist

Hobby: Singing, Listening to music, Rafting, LARPing, Gardening, Quilting, Rappelling

Introduction: My name is Foster Heidenreich CPA, I am a delightful, quaint, glorious, quaint, faithful, enchanting, fine person who loves writing and wants to share my knowledge and understanding with you.