Originally we envisioned using Objectivity as the low-level DBMS in the WSA, echoing the implementation of the SDSS-EDR. However, due to some performance issues experienced by the SDSS science archive team at JHU, we are following the SDSS group and investigating Microsoft `SQLServer' (a relational DBMS). We currently envision that the WSA design will therefore be based on SQLServer, but we note that we are not inflexible as to the final DBMS choice - we also intend to investigate the IBM's `DB2' (again, a relational DBMS).
Feasibility of DBMS choice will come down to performance of the implementation of the SSS archive, which is similar in size to the projected size of the WSA. This work is being undertaken at the time of writing. From the software point of view, we do not anticipate any implementation problems for Tbyte-sized archives.