We are in the middle of a design process, but expect to freeze implementation decisions quite soon. A Science Requirements Document, specifying both data product definition and archive functionality, is being developed with JAC and the UKIDSS consortium and should be agreed by the end of 2002. We have been experimenting with a variety of DBMS packages, and will make a final choice in March 2003. Our working assumption is that we will follow developments by the SDSS team in order to save development effort and also to simplify UKIDSS - SDSS cross matching. This will probably mean using MS SQL Server and an updated equivalent of the SDSS query tool. We expect to do further development in parallelisation of SQL server, and an improved user interface. On the hardware side, we will be making choices by December 2002. It is likely that we will implement storage and search engine simultaneously with a cluster of roughly 50 nodes, in combination with a single large RAID array. (We may also have backup nearline storage such as DVD towers or a tape robot). In addition we expect that our analysis engine will be a share of a high powered multi-processor SMP machine, through a collaboration with the UK National e-Science Centre (NeSC), which is a joint venture of Edinburgh and Glasgow Universities. This variety of hardware will by default hang off our Local Area Network, but we have also been in discussion with some vendors over the possibility of a more integrated approach using a dedicated fibre-channel SAN. As well as the obvious cost issue, a key driver in our final hardware decisions is the requirement to find a solution that is (affordably) scaleable to VISTA and beyond.