I have had the situation recently where I was working remotely over the VPN on a data link that left a lot to be desired. Unfortunately my area isn’t serviced that well by ADSL and the NBN is years away.

Opening and saving back to a central file of about 50mb wasn’t too much trouble, but opening the 400mb of links was a massive problem. I needed a solution so I could work on my files and fast.

What I ended up doing was using the subst command in DOS to generate a pseudo network drive on my local system and replicated the folder structure below that. The central file and all of the associated links were saved at their UNC path \\work-network\projects\job number\revit

I opened the central without it’s links and then repathed the links to their windows path z:\job number\revit

The result being that I could access and save to the same central file as the modellers back in the office and no one was affected by linked models being pathed to the wrong place because on their next save to central / reload latest, the paths of the links were updated to z:\job number\revit, everything worked seamlessly.

Of course, a few times I forgot to subst the drive before loading the central file which obviously resulted in links not loading, so to solve this problem I put together some code in c# which performed the drive subst for me and then set the program to run on startup.

The code is quite basic, really only a single line

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.ComponentModel;
using System.Runtime.InteropServices;

namespace CreateVirtualDrive
{
 class Program
 {
 static void Main(string[] args)
 {
 System.Diagnostics.Process.Start("cmd.exe", "/c subst z: d:\\work");
 }
 }
}

Just keep in mind that this is only generating the drive letter, not the directory structure, so you need to ensure that the directory structure between the pseudo drive and the server match exactly otherwise it will not work.

It’s not a perfect solution, but it works well with minimal effect to everyday workflow. It does exactly as intended, eliminating the wait time loading large chunks of data across a slow VPN link. The linked models only need to be copied across the VPN when updated rather than on each open.