RAL Antares (CTA) plans with TAPE REST API

LHCb already accessing RAL Antares with WebDAV protocol
- this protocol currently doesn't provide features for full tape system management
- as a workaround xroot is currently used to fill this gap
- sub-optimal multi-hop data transfers through Echo
  - sufficient for expected low 2022 data rates
  - RAL would like to transfer data directly for 2023 full physics run
ATLAS & CMS currently use Antares with xroot protocol
- practical point of view - use common protocol (HTTP-TPC) everywhere
- these experiments could also benefit from early TAPE REST adoption at RAL

TAPE REST API timeline

CERN Plan: Presentation for CTA+FTS in the attached slides
RAL plan
- have the production endpoint with TAPE REST API by February 2023
- everything ready for 2023 Run (~ April)
Both plans seems to have compatible timeline and looks realistic
ATLAS use additional transfer metadata
- must be passed from Rucio via FTS to TAPE REST API
- some development necessary in Rucio for ATLAS
- not used by other experiments
Rough ATLAS plan for replacing SRM with TAPE REST
- winter shutdown 2023/24 seems to be good opportunity to ask T1 for dCache upgrade => TAPE REST support
- it is technically possible to start with early / active adopters of TAPE REST sooner (e.g. with RAL)
  - sites can be moved one by one once they become ready

FTS will provide the tests to validate the REST API for the different endpoints.
We (as BDT team) need to encourage the endpoints to join and complete the matrix.
we would like to have
- CERN devel CTA - already deployed
- FNAL dCache with their implementation
- RAL will also provide testbed because they are planning to be early adopters
added TAPE REST API twiki page to document details

lightweigh - no changes in file transfers, just different call to bring file online
REST API functionality tested by developers
FTS functionality / Dirac direct interaction with TAPE endpoints validated together with experiments