Skip to content

Latest commit

 

History

History
179 lines (134 loc) · 5.28 KB

draft-agntcy-ads.md

File metadata and controls

179 lines (134 loc) · 5.28 KB

Description of Agent Directory Service

title: "Agent Directory Service" abbrev: "agent-dir" category: info

docname: draft-agntcy-ads-latest submissiontype: independent number: date: consensus: false v: 3 area: Applications workgroup: Independent Submission keyword:

  • AI
  • Agentic AI
  • Directory Protocol
  • Agent Discovery
  • DHT venue: group: WG type: Working Group github: agntcy/dir latest: https://spec.dir.agncty.org

author:

fullname: Luca Muscariello
organization: Cisco
email: [email protected]

normative: RFC6920:

informative: DHT: DOI.10.5555/646334.687801

--- abstract

The Agent Directory Service (ADS) is a distributed directory service designed to store metadata for AI agent applications. This metadata, stored as directory records, enables the discovery of agent applications with specific skills for solving various problems. The implementation features distributed directories that interconnect through a content-routing protocol.

--- middle

Conventions and Definitions

{::boilerplate bcp14-tagged}

Introduction

The Agent Directory Service (ADS) provides a mechanism for discovering AI agent applications based on their capabilities. It uses a distributed directory architecture to map agent skills to directory record identifiers and maintains a list of directory servers hosting those records.

Directory records are identified by globally unique names that are routable within a DHT (Distributed Hash Table) to locate peer directory servers. The skill taxonomy is also routable in the DHT to map skillsets to records that announce those skills.

Each directory record includes skills from a defined taxonomy, as specified in the [Taxonomy of AI Agent Skills] from [OASF]. While all record data is modeled using [OASF], only skills are leveraged for content routing in the distributed network of directory servers.

This document describes the core concepts and architecture of the Agent Directory Service.

Naming

In distributed systems, a reliable and collision-resistant naming scheme is crucial. The agent directory uses cryptographic hashes {{!RFC7838}} to generate globally unique identifiers for data records.

ADS leverages OCI as object storage, and therefore identifiers are made available as described in [OCI digest].

Content Routing

ADS implements capability-based record discovery through a hierarchical skill taxonomy. This architecture enables:

Capability Announcement

Multi-agent systems publish their capabilities by encoding them as skill taxonomies. Each record contains metadata describing the agent's functional abilities. Skills are structured in a hierarchical format for efficient matching.

Discovery Process

The system performs a two-phase discovery operation:

  1. Matches queried capabilities against the skill taxonomy to determine records by their identifier
  2. Identifies the server nodes storing relevant records.

Distributed Resolution

Local nodes execute targeted retrievals based on:

  1. Skill matching results: Evaluates capability requirements.
  2. Server location information: Determines optimal data sources.

ADS uses libp2p [Kad-DHT] {{DHT}} for server and content discovery.

                             +----------------+
                             |    DHT Node    |
                             | Content Index  |
                             +----------------+
                                    ^
                                    |
                   +----------------+-----------------+
                   |                |                |
           +-------v------+  +------v-------+  +-----v--------+
           | Server Node A |  | Server Node B|  | Server Node C|
           |   Content X  |  |   Content Y  |  |   Content Z  |
           +-------+------+  +------+-------+  +------+-------+
                   |               |                  |
                   |        Content Exchange          |
                   +---------------+------------------+
                                  |
                          Content Replication

Flow:
1. Servers register content with DHT
2. DHT maintains content-to-server mappings
3. Servers query DHT to locate content
4. DHT returns list of servers hosting content
5. Servers download content from peers

Distributed Object Storage

ADS differs from block storage systems like [IPFS] in its approach to distributed object storage.

Simplified Content Retrieval

  1. ADS directly stores complete records rather than splitting them into blocks.
  2. No special optimizations are needed for retrieving content from multiple sources.
  3. Records are retrieved as complete units using standard OCI protocols.

OCI Integration

ADS leverages the OCI distribution specification for content storage and retrieval:

  1. Records are stored and transferred using OCI artifacts.
  2. Any OCI distribution-compliant server can participate in the network.
  3. Servers retrieve records directly from each other using standard OCI protocols.

While ADS uses zot as its reference OCI server implementation, the system works with any server that implements the OCI distribution specification.

IANA Considerations

This document has no IANA actions.

--- bac

Acknowledgments

{:numbered="false"}