"J.W. Bizzaro" wrote: > I also think the BL needs to handle the management of permanent network/volume > storage. No problem. I think I hardly need to change anything about the storage, only need to realize the retreival code. GMS currently stores all data that passed its net to disk. (or gms naming: 'historyIO') This is kinda rough, but in principal it's what you need. > > > The DL/UI will handle the temporary loading of networks/volumes into a > workspace. Yep, by doing so reducing processing load. > > > UI UI > | | > \ | / \ | / > rootDL DL > | | > \ | / \ | / > BL > > > As per Jarl's suggestion. So, there really is only DL->DL authentication, > correct? No, arf, this is my suggestion: 1) 2) 3) UI UI UI | | | \ | / \ | / \ | / rootDL DL <--- DL | | \ | / \ | / BROKERING LAYER 1- rootDL log AUTOMATICALLY into the BL, 2- this rootDL can grant other DL's access to the BL, 3- every DL can deside to grant other DL access to ITS OWN nodes. (but maybe the rootDL shouldn't be able to do (3)? ) Clear? bye, jarl