Commavee

Moments ago I finished a rather complex set of connections that allows me to use all of my systems exactly the way I wish.

After ridding my network of the Windows boxes and the RG2000 router, I was faced with creating interconnections to my new Apple gear. I don’t mean connections from the Apple gear to anything else - I mean connections from my existing network devices, such as my KVM and other router to the Apple gear.

Over the years, my home network has morphed into a hybrid of personal computing devices (PCs, printers, scanners. etc.), servers and routers. This includes all of the equipment I use to run freepository, my consulting business, develop software, do graphics work and write. I run all of my development and production systems on Linux (FC3). As my computing needs evolved, I became overwhelmed with each device’s keyboard, mouse and monitor. No surprise, I added a big KVM switch about five years ago. This reduced the KVM mess to a single set of input devices on my desk. I could now hot-key between my development server, production server, and business laptop.

Nice… but I needed more.

Three years ago I added a second physical monitor to the mix, using it from my primary development / desktop system. I exported the display of other systems to it using either x2vnc or x2x (both are great, but x2x is limited to just that X->X, whereas x2vnc successfully attaches to any vnc server). This allows me to have the monitors side-by-side on my real desk, creating a virtual mega-desktop that is divided between the two. When I drag my mouse across the west display, it appears to leap across air to the edge of the next monitor, which is displaying the desktop of my exported system. This has allowed me to increase my productivity, literally working on two systems almost simultaneously. This is in addition to hot-keying to the other systems on the KVM.

Once again - nice…. but…

When I replaced the old Windows gear, I needed to replicate the same functionality on my Apple equipment. I didn’t expect this to be a big deal, but the Airport Extreme added some complexity. It can successfully bridge its wired and wireless ethernet ports (necessary to allow resources on either segment to be seen and used by any client served by the Airport) - but - and this is a BIG BUT - the addressing sub-nets for the wired interface and the range served by the wireless DHCP server must be the same, if the wired interface’s address is set manually. If the wired interface’s address is set via DHCP, this restriction doesn’t apply and the bridging functions as expected.

That is, if the DHCP server’s address range is 10.0.1.x - 10.0.1.y, then in order for the bridge to *really* be effective (as opposed to simply providing one-way NATed resources), then the wired side of the Airport Extreme must be set to a 10.0.1.x address also. That wouldn’t be bad, so long as everything was on the private subnet. But I don’t have my network setup like that, and apparently this is a well-known quirk of the Airport Extreme.

I resolved the issue after getting a suggestion from one of my networking friends who runs a large ISP. He suggested setting up the Airport Extreme to use DHCP for its wired side address & keeping the private wireless address as-is. So I setup another DHCP server to serve only one address - a wired subnet address to the Airport Extreme. This satisfied the Airport Extreme’s rules and allowed me to control the Powerbook while it is connected to the wired network using the existing hardware setup.

All that was left was to install OSXvnc on the Powerbook, export the display and presto - I’m now typing this post on the Powerbook through the exported display that is adjacent to the first of the two physical monitors, using the unified mouse and keyboard.

Nice. Very nice.