So I've got the Sun Ray up and running. I decided to put the noisy Sun w2100z workstation in my closet. In all fairness, once I got it in there, I couldn't hear it over the desk fan I have running 24/7 in there to push air about. I also asked my wife, working on a puzzle 5 feet away about the noise and she asked, "What noise?"

I did close the door 3/4 of the way and when I went in later, it was hot. I put it wide open (which is normal) and when I came back later, the heat was okay.

Here is a snap of my closet gear:

So we put in a bathroom ceiling fan into the space above the garage (the closet is on the second floor). You can catch a glimpse of it by the rug standing up. I also put in a desk fan to push air out the door. The blue thing the computers are standing on is a 3/4 width telco rack. A filer didn't fit in it, so I've never used it as anything other than a table. I've got a UPS device inside it. The network hub is a Gig Netgear box.

You can also see the Livingston Portmaster (darkness) I scavanged from a lab cleanup. Both the Ultra 10 (ultralord) and the Ultra 5 (sandman) have working consoles on darkness. There is also a small patch panel by the fan. Currently one cable goes to the hub and the other carries console back for the Cisco 831 in my office.

Hmm, as normal, I have Shuttle skins sitting about. The Shuttle SS51G was a disappointment as far as heat and noise went for me. The PSU fan is the loudest component. I replaced one and the PC was silent. But travel took its toil eventually and the MB got fried by a screw which worked its way loose. Also, I had another one skinless and placed a composition notebook on top of the exposed harddrive. After 3 days of this, the HD died. The notebook was hot to the touch.

I typically also have a SS51G in this closet.

I have another closet in my office which I used to use, but when I got tired of the filer I had, a F740, I put the stuff back in this closet. The filer eventually got sent to NetApp NGS in RTP to do with as they saw fit.

If you squint real hard, you can see the config info in pencil on top of the Portmaster. Heck, if you get good enough image processing software, you could have root on it. Hah! I had to go check to make sure I wasn't violating local security efforts - there is no password present.

Update 1: As a followup, the Sun Ray 1G is connected to a Cisco 831, which in turn is connected to Sun directly. I forgot to mention that once I get the console working on the Sun w2100z, I'll VPN in with it and be able to connect to the Sun w2100z via the Sun Ray 1G. Yucky, but it will allow me to have two machines on SWAN and also allow me, as Spencer calls it, to have "The Holy Grail of Local Builds".

I'm not blaming my Aunt Patti, but it was only after I explained to her what a Bakeoff was that we had to change the name. I was visiting Ann Arbor for my first Bakeoff and while she knew I worked with computers, she thought I had come up to cook. At some point, Pillsbury had some words with someone and Bakeoff became Bakeathon. I'm sure it all started with an innocent phone call where my aunt was asking them how I did at the competition.

Seriously, I don't live in the Valley and I sometimes have a hard time explaining to people what it is that I do and what I do when I go to these events. I thought I'd try and explain them here:

• Connectathon is an annual event where multiple protocol vendors come together to test interoperability. While it sounds like a competition, most developers check that attitude at the door. Everyone signs a NDA and we only allow developers in to the event. Typically we might have a 1-2 hour Open House, but otherwise if we find someone from Marketing, we string them up. This is a large event and multiple protocols will be tested.
• Bakeathons are less organized and typically focus only on NFSv4. Although sometimes we might let NFS over RDMA in for some testing. Again, NDAs are in effect, but the testing and atmosphere is much less formal. Most Marketing types do not want to be stuck in a confinded and sweaty space with technical types ignoring them.

In general, the attitude is not so much I can find N bugs in your implementation as opposed to together we can strengthen the reliability in the protocol and attract more customers. Kinda an innocent assumption that the pie is big enough to share. A more realistic one might be we convince everyone to become dependent on our protocols and then we fight for market share. :->

You'll also see cutting edge code at these events - i.e., unit testing has not been thought of yet. But it is only as multiple vendors butt heads that a consensus is arrived at and the protocol changes.

One of the goals with the IETF and the NFSv4 community is that there are 2 or more implementations of new protocol extensions before they sign off on them.

One thing I do find amazing at these events is that you typically find developers and QA testers. One would think you would have QA engineers banging away at different boxes and feeding the bugs to the developers. These engineers are quite busy churning out new code. But no, if the code is not robust enough to get unit testing, it isn't good enough to let loose the QA engineers on.