DTE Infrastructure Component

Teapot

Federated Data Infrastructure

Description

An easy-to-install edge service that a facility may deploy to provide remote access to their storage, in order to facilitate data ingress and egress.

Teapot is an easy-to-install edge service that a facility may deploy to provide remote access to their storage, in order to facilitate data ingress and egress. It provides the reference implementation of the teapot architecture: a scalable and technology agnostic approach to support data ingress and egress.

Unlike existing storage solutions that are compatible with Datalake, teapot does not require exclusive control of the underlying storage. Instead, it supports existing access methods. This allows facility admins to deploy teapot on top of existing storage solutions without disrupting established modes of access.

Teapot is a core component that most users do not access directly.  Instead, teapot is used by any DT manager or user who is (directly or otherwise) managing data locality.

One reason for transferring data is to satisfy the data requirements of some workflow when the required data is missing from the local storage of the facility that will run the workflow.  Under these circumstances, the user that triggered the workflow is using the teapot software if either the source facility (hosting the data) or destination facility (requiring the data) has deployed the teapot software.

Release Notes

This release represents a technical preview of the interTwin federated data management solution.

There are two external software components: FTS and Rucio.  They are fully established projects, independent of the interTwin project.  The software is production-ready, at TRL 9, and hardened with many years of production-critical use.  Both projects have multiple deployments of their software, operated by different communities.

The ALISE software is currently in a development phase, under the aegis of interTwin.  At the time of release, ALISE is TRL 4.  The user-facing functionality of ALISE is mostly feature-complete; however, anticipated changes to the API imply that the necessary integration work (whereby a service uses ALISE to identify a user) should be considered experimental. Feedback from early adopters is encouraged, but any plans to deploy ALISE should be tempered by the anticipated changes to the API.

The teapot software is also currently in a development phase, within the interTwin project.  Teapot is TRL 3–4.  The current release is sufficient to build a limited proof-of-concept demonstration, supporting the data transfer requirements of a single user.

Future Plans

A number of improvements are planned for teapot. Teapot will be updated to support multiple, concurrent users. The per-user WebDAV instance management will be automated, starting new services on demand, and terminating them if there is sufficient idle time. It will also be updated to integrate with ALISE, to support automated identity management.

For ALISE, we anticipate possible improvements and stabilisation of the service-integration API, based on experience gained from integrating ALISE into various services. In addition, we plan to add support for client authentication in future versions of ALISE. This will limit access to the identity mapping information, providing this information to authorised services only.

There are currently no immediate plans to improve FTS. Future experience from integrating the testbed with the interTwin service providers may identify necessary enhancements to FTS.

Plan to enhance Rucio by improving support for integrating Rucio with external catalogues; e.g., Lattice QCD’s catalogue.  Other plans to enhance Rucio may emerge from our experience when integrating Rucio with DTE core service, DTE thematic modules, and with the science use cases.

 

Target Audience
+

Teapot is a core component that most users do not access directly.  Instead, teapot is used by any DT manager or user who is (directly or otherwise) managing data locality.

License
+

Apache 2.0

Created by
+