1439: Rack Unit

Explain xkcd: It's 'cause you're dumb.
Jump to: navigation, search
Rack Unit
There's also nothing in the TOSes that says you can't let a dog play baseball in the server room!
Title text: There's also nothing in the TOSes that says you can't let a dog play baseball in the server room!

Explanation[edit]

Black Hat announces to Megan that 19-inch racks for datacenter servers and Langstroth hive frames are both 19 inches wide (482.6 mm), with similar spacing between each slot. Black Hat is motivated by this knowledge to break into a Google datacenter and fill server racks with beehives. When Megan sarcastically consoles Black Hat for the loss of his hives, he declares that he'll find other datacenters to install hives in.

The pitch (or distance between repeating items) of 19 inch rack server hardware is measured in rack units (U), and is standardized at 1.75" (44.45 mm). Langstroth frames are typically mounted at a pitch of 1.5" (38.1 mm), and as a result would fit relatively well with a server cabinet. In contrast to the horizontal orientation of the modules in a server rack, honeycomb frames are designed to hang vertically, so the cells can hold nectar without it dripping out. How Black Hat was able to re-orient the racks to suit the needs of honey production remains a mystery.

Some datacenters provide colocation services where customers may install a server at a central location with better bandwidth and power reliability than a customer could provide on their own. Noticing that typical colocation terms of service (TOS) agreements don't specifically rule out the installation of beehives, Black Hat suggests he can enter a legal contract allowing him to install beehives at a data center without being kicked out.

As with many types of terms of service contracts, such as privacy agreements, a standard boilerplate wording is used amongst many contractors. Megan comments that a term forbidding the installation of beehives will quickly become a standard term in colocation agreements. The comic thus might be making fun of boilerplate legal agreements, which are commonly signed without being read, suggesting that the legal wording is so long and detailed that anything not specifically prohibited must be allowed.

The title text is a reference to the film Air Bud. The original quote is "Ain't no rules says a dog can't play basketball."

Transcript[edit]

[Megan and Black Hat are talking, Megan is sitting at a computer.]
Black Hat: I've discovered something.
Megan: Oh?
Black Hat: Standard server rack units and standard beehive honeycomb frames are compatible.
Black Hat: They're both 19 inches, with similar pitches.
Megan: Uh oh.
Black Hat: I'm pleased to announce that today, for a few hours, Google led the world in datacenter honey production.
Black Hat: Until their security people kicked me out.
Megan: I'm sorry your beekeeping career ended so quickly.
Black Hat: I'll find a new datacenter.
Black Hat: Turns out most colocation TOSes don't mention beehives.
Megan: I suspect that will soon change.


comment.png add a comment! ⋅ comment.png add a topic (use sparingly)! ⋅ Icons-mini-action refresh blue.gif refresh comments!

Discussion

Bzzzzz 108.162.250.202 04:52, 27 October 2014 (UTC)

Bzz ZZ zz bz?! (What did you say about my mother?!) 108.162.216.26 19:09, 27 October 2014 (UTC)
Bzz, zzz bzzz bz. Bzz zz. 108.162.250.202 23:30, 27 October 2014 (UTC)
bzbbzbbb bzzbbzbz bzzbzzbb bzzbzzbb bzzbzzzz bbzbbbbb bzzzbzbb bzzbzbbb bzzbbzbz bzzzbbzb bzzbbzbz Mikemk (talk) 05:31, 27 December 2016 (UTC)

Air Bud has had multiple mentions in his comics, but I don't know all of them. I also think it was mentioned in one of the What-If's. I'll do a quick Google search to see if I can get at least one of them. 108.162.237.161 06:55, 27 October 2014 (UTC)

Mouse-over text in the final image. http://what-if.xkcd.com/111/ Still searching. 108.162.237.161 06:58, 27 October 2014 (UTC)
http://xkcd.com/115/ 108.162.237.161 07:00, 27 October 2014 (UTC)

This comic may be a reference to the highly hyped lack-rack https://wiki.eth0.nl/index.php/LackRack --Belibem (talk) 09:43, 27 October 2014 (UTC)

It's unlikely that this was intentional, but this comic is almost the same as this panel from MSPA: http://www.mspaintadventures.com/?s=6&p=003976 108.162.219.224 11:50, 27 October 2014 (UTC)

I'm sure an apiarist might know whether there'd be a problem with air temperature (often chilled), on any bee colony. And with the lack of non-plastic internal flora, and almost certainly some quite severe filtration screens betwixt server environment and the outside world I can't see [i]any[/i] chance for nectar collection. Of course, it's Black Hat, so he's probably worked around both of these (slotted in next to blade servers with consistently warming processors, and maybe a ready nectar supply. But I stil feel for the poor bees, with all those whirring fans of all shapes and sizes, around and within the server room equipment. (Also, perhaps interesting to note that apparently most colocatiopn TOSs don't mention beehives. So he found that some did...) 141.101.99.112 12:47, 27 October 2014 (UTC)

I'd think that a more-or-less simple restriction on interference with other tenants and their equipment would be sufficient to proscribe Black-Hat from causing any issues with his bees (officially, at least), wouldn't it? So perhaps it's as simple as a TOS that only proscribes electromagnetic interference (maybe even touching equipment of other tenants), but nothing involving bees per se? -- Brettpeirce (talk) 13:11, 27 October 2014 (UTC)
Lack of non-plastic flora might not be a problem, since honeybees are opportunists, and will gather any sugary liquid they can find. Someone once told me that the bees in the Smithsonian's observation hive made honey from what they found in discarded beverage containers left around the mall by human pedestrians. 173.245.54.153 01:11, 28 October 2014 (UTC)

What does "similar pitches" in the comic mean? 108.162.216.73

not entirely sure about an industry-general term, but there is mention of "tile pitch" here: http://www.dell.com/downloads/global/products/pedge/en/energy-smart-containment-rack-deployment-guide-dell.pdf -- Brettpeirce (talk) 14:41, 27 October 2014 (UTC)
Ah, here's something I found: "Aisle pitch is the distance from the center of one cold aisle to the center of the next cold aisle either to the left or right. Data centers often use a seven-tile aisle pitch. This measurement allows two 2 x 2 foot (0.61 x 0.61 m) floor tiles in the cold aisle, 3 feet (0.9 m) in the hot aisle, and a 42-inch (1-m) allowance for the depth of the cabinet or rack."
...and though it doesn't seem the term "pitch" is used in bee keeping that I can find, there are probably generally followed guidelines on spacing... http://www.tillysnest.com/2012/04/placement-of-beehives.html -- Brettpeirce (talk) 14:49, 27 October 2014 (UTC)
Per wikipedia -- " "Pitch" is widely used to describe the distance between repeated elements in a structure possessing translational symmetry". This would include things like server racks (the distance from the bottom of one slot to the bottom of the next), beehives (the distance from one pane to the next). You commonly hear it in relation to airline seats ("seat pitch" -- the distance from one seat to the next, as a measure of the relative comfort of airline seats) but it is a more general term. Vyzen (talk) 15:13, 27 October 2014 (UTC)
Engineer here - The term pitch is used to describe the distance between repeating items. For example: The holes have a 10mm pitch = There is a hole every 10mm. I've added the pitch spec for server racks and beehives, which are not too dissimilar. --Pudder (talk) 15:27, 27 October 2014 (UTC)
Without looking it up (sorry, feeling lazy), it occurs to me that the term pitch might ultimately have derived from screwthreads. Pitching (especially when it comes to ships, and by extension air and space-craft) regards a sloping angle, and the angle of the thread dictates the linear distance between each circuit of the ridge/thread element. This latter property is more handily measured than the perpendicular angle away from the pure tangent (assuming not multi-threaded, although that's rarer in nuts and bolts, etc, due to less inherent mechanical advantage), and so while the cutting may have been dictated by the angle (i.e. pitch), the definition quickly becomes standardised against the linear periodicity, and thus becomes used even in describing perfectly square measurements, such as screwholes in racking. Just an idle thought. That adds nothing to the discussion. 141.101.99.112 21:55, 27 October 2014 (UTC)
Actually, the pitch of the frames in a Langstroth hive is variable. Some beekeepers use ten frames to a super (box) and others use nine or eight, spacing them by eye. With too much space between frames, the bees tend to build bridge comb (a form of undesired "burr comb") between frames, making it difficult to work the hive. Frames sometimes need to be taken out, mostly for inspection or honey harvesting. "Bee space" is a familiar concept to every modern beekeeper. Much less than 3/16" between components, and the bees will seal that space shut with bee glue, or propolis. Much more than 3/8" and they will fill the vacancy with wild comb, or burr comb. Spaces dimensioned between those limits are left clear for bees to move around in. I'm having fun imagining going into a server rack with a smoker and hive tool (a little pry bar to separate the supers and unglue the frames from their support.)
Orienting the frames horizontally in a normally situated rack would be a deal-breaker, though. Honeycomb cells are built with their bases on a vertical foundation, with each cell having a slight upward tilt, on the order of 10° to 12°, so that the nectar doesn't fall out. The bees fan their wings to ventilate the hive, reducing the nectar's water content and increasing its viscosity, but they also manage the ventilation to maintain a certain warmth around the brood comb. At that temperature, honey flows pretty well. 173.245.54.153 01:28, 28 October 2014 (UTC)
This is my first post in this wiki. 1.) the difference between 1.5" and 1.75" is enough to let an inserted object give the ~10 degrees necessary for the honey to not drip out. 2.) In the data center I'm familiar with the server blades were mounted vertically (Verari BladeRack 2) would vertically aligned boxes for for bee keeping? 199.27.128.182 00:33, 2 November 2014 (UTC)
Here is a picture of a frame of honeycomb lifted out of a super. In the beekeeper's right hand you can see the bit of extended top bar that rests on a rabbet in the top edge of the box. Here is a picture of what beekeepers call the woodware or bee furniture. That is what it looks like before the bees have used it and covered it with wax and propolis. That tan sheet in the lifted frame is beeswax foundation, embossed with a hex pattern of worker-sized cells. As the bees draw comb out from both sides of that foundation, they instinctively give the cells their upward tilt. (Here is an image with the tilt angle exaggerated.) No need for a human to tilt the frame for them. You can see how the top bars in a beehive are horizontal, with the frames hanging vertically down from them. In the Verari box, the "top bar" of each blade is vertical. We might be able to work with that... read on:
Frames start out with their top bars resting on rabbets in the front and back walls of the super, which is a topless, bottomless box, one module of a beehive. As time goes on, the bees cover the inside surfaces of their nest with propolis, which sort of glues the frames into the box. The burr comb, that you can see on the top and bottom bars in the first picture, gets broken when the frame comes out. Honey-filled burr comb scraped off the top bar makes a nice snack, tucked under the veil on a sunny afternoon. The yellow thing in that first picture looks like the handle of a hive tool, used to scrape off unwanted bits of comb, pry the frames loose, and get some finger room under the ends for lifting. When the frames go back in, it's just gravity holding them in place. This picture shows some bridge comb between frames. A careful stab with the hive tool will cut that, so it doesn't tear as the frame comes out.
Slotting and unslotting a frame with its top bar oriented vertically would call for some dexterity. The need to secure the ears to the rack might call for an extra pair of hands. If the frame wiggled too much to one side or the other on its way in or out of the rack, some bees might get crushed... when that happens, they release an alarm pheromone (which smells like banana oil, or isoamyl acetate, familiar to shooters as Hoppe's #9 nitro solvent.) Other bees get "defensive" when they smell that, and come looking for animals to sting. When they get really grumpy, they launch straight for the face, which is why we wear veils made out of metal screen.
Whew, that's enough typing for now. I will check in on this page now and then. If I've left anything unclear, ask away! 173.245.54.153 02:54, 2 November 2014 (UTC)
Thanks for your insight, I found it very interesting. I never thought I would be learning about bees today! As far as the frame pitch, my research suggested that there wasn't any 'standard' pitch, but that 1.5" was fairly typical. I would certainly defer to your greater expertise if you feel the article needs editing.--Pudder (talk) 08:41, 28 October 2014 (UTC)
I do not know what resources Black Hat can command. I can only assume the racks were somehow turned on their sides, so the bees could work in their customary orientation with respect to gravity. Sensitive little darlings they are, and they smell nice. I've put a parenthesis in the explanation, more or less to that effect. 173.245.54.153 00:28, 30 October 2014 (UTC)

I wouldn't bee suprised if this became of those 'reality imitates art' situations, and somebody goes and makes a beehive out of an old server cabinet. I'll just leave this eBay auction here.... --Pudder (talk) 15:31, 27 October 2014 (UTC)

I'd be cautious if I were to do this. This is one of those projects that could develop some serious bugs... 173.245.54.202 (talk) (please sign your comments with ~~~~)

This is certainly related to the famous ant-farm-computer, Hex, invented by Terry Pratchett. Anthill Inside! ::: 108.162.249.231 02:48, 28 October 2014 (UTC)

That certainly is one way to run a honeypot.--Henke37 (talk) 14:49, 28 October 2014 (UTC)

I remember seeing an article a few years ago about a company planning to use sideways server racks for liquid immersion cooling. I wonder if honey is electrically or thermally conductive... http://hardware.slashdot.org/story/10/03/18/1955238/Startups-Submerged-Servers-Could-Cut-Cooling-Costs 108.162.216.65 15:54, 29 October 2014 (UTC)

Honey is a supersaturated sugar solution made from a much more dilute sugar solution, usually nectar. Not a lot of ionic species in the solution, so maybe slightly more electrically conductive than DI water. The transformation from nectar to honey involves controlled airflow, from bees fanning their wings while remaining stationary on a wood or wax substrate that they can hook their feet onto. Liquid coolant would not be compatible with that process, to say nothing of the air-breathing insects doing the work.
If the bees chose to draw comb on a foundation of active circuit board, the comb would act as thermal insulation, and would interfere with air flow around the hot semiconductors or their heat sinks. Data-center honey production would call for intense hive management to avoid issues like that. If the honey frames were not kept separate from the electronic blades, the frames would need to be pulled on a rotation on the order of four times daily, to check the electronic modules and scrape them off if needed. Interesting work, if you can get it. 173.245.54.153 21:22, 30 October 2014 (UTC)

Many servers use hypervisors for virtualization. What's interesting is that there is a hypervisor called bhyve, pronounced as "beehive". 103.22.201.128 (talk) (please sign your comments with ~~~~)

There's another Air Bud reference in 1552:_Rulebook AmbroseChapel (talk) 04:21, 24 August 2017 (UTC)