Hi Mathew
What are you comparing this to ? another Oracle installation ? SQL Server ? a previous version of EnterpriseOne ?
Improving the performance of a package build can be through a number of different paths. Remember, performance of something like this could be held up by either the client, the network or the server (ie, everything) - but it is the lowest speed device that reduces everything.
So, while a package build is occurring, you need to look at several things.
1. How is the client performance (the deployment server) ? How is the CPU ? Disk speed ?
2. Have you tried running a package build using a fat client (development workstation) - did it dramatically change the performance (fast/slower ?)
3. While the package build is occurring - how is the performance of the database server - CPU ? Disk utilization ?
4. What is the throughput on the network ? This is an easy one to test, even before you run a package build - try pinging the enterprise server from the deployment server (and vice versa) - and note the latency. Try increasing the PING size to 32k to see if there is any substantial difference. You should see <1ms of latency for small packets. THEN, try FTP'ing a file between your enterprise server and deployment server. Try FTP'ing a large file - about 2Gb or so - and calculate how long it is taking to transfer.
Because you're using Oracle as a database, my guess is that you're on Unix - and there might be an issue with your network performance - something might not be set correctly - and with large transfers, you're probably seeing that issue magnified (but believe me, it would also be occurring on your day to day operations).
I recently worked with a customer who had just purchased a new Unix platform. Evidently, there were some changes on this new platform that prevented large transfers of a single thread across the network. The manufacturer pointed out that the box was designed as a "web server" - hence it couldn't necessarily handle the types of transactions that JDE requires without substantially changing the architecture. I'm not mentioning platform names - but often the hardware manufacturers make changes to keep their costs down - and these become detrimental to large ERP system configurations...