knightfoo
2005-04-01, 11:50 AM
Nearly 2 years ago ServerBeach announced a revolutionary new tool for dedicated server customers: RapidReboot(tm), the first datacenter-wide customer reboot control solution. Customers and competitors alike have wracked their brains since the release of the RapidReboot(tm) technology, trying to figure out just how the technology works so well and still be competitively priced in the discount market. As of today, no one has any success in cracking the secret of RapidReboot(tm).
Many providers have released their own reboot systems, but there is always a catch: you have pay for it, only certain customers can use it, it only works between the hours of 9-5 M-F, etc. Not RapidReboot. All customers in all datacenters can use RapidReboot 365/24. Since RapidReboot(tm) has been public for two years, and we are confident that no one could duplicate the technology even if they knew what it was, here it is: DC Monkeys.
What is a DC Monkey? A Rhesus monkey which is specially trained to quickly identify servers that need to be rebooted and get to them in a matter of seconds, regardless of datacenter location. After many years of studying datacenter operations, we found that the most time-consuming part of a server reboot was locating and getting to the server. Long rows of contiguous racks and servers nearly identical in appearance make it difficult for humans to easily navigate and find servers. However, Rhesus monkeys are much more agile and can cope with "forests" of racks, easily leaping over and swinging around rack poles to reach remote servers.
The monkeys are trained to locate servers in two stages. First, they are shown a map of the datacenter with the rack affected highlighted. Since the monkeys are normally above the racks, its very natural for them to utilize this "look down" of the datacenter to find the rack location. In the second stage, they are shown a representation of a rack with the relative locations of the servers in need of a reboot highlighted (picture attached). A fully-trained monkey can identify a rack and server location in under 2 seconds. After identification, it only takes about 5 seconds for the monkey to reach the rack and toggle the power switch on the proper server. This is about as much time as it takes for a browser page to reload .. nearly instant!
A datacenter with 5,000 servers can be handled by a team of 12-16 monkeys, even on a very busy day. They work on rotating shifts with plenty of resting time between reboot shifts. Rhesus monkeys are very mellow and not easily distracted, which makes them ideal for this sort of work. Each datacenter has a jungle atrium built-in to allow the monkeys some "time in the wild". Even though we have a flawless safety record to date, we have on-site zoologists and doctors just in case.
How cost-effective is this solution? Very! Even with a team of 12 monkeys in every datacenter, and the costs associated with caring for them, the cost per server is only $10 per month. This allows us to provide the technology to all of our customers at no cost and still maintain a healthy profit margin.
Many providers have released their own reboot systems, but there is always a catch: you have pay for it, only certain customers can use it, it only works between the hours of 9-5 M-F, etc. Not RapidReboot. All customers in all datacenters can use RapidReboot 365/24. Since RapidReboot(tm) has been public for two years, and we are confident that no one could duplicate the technology even if they knew what it was, here it is: DC Monkeys.
What is a DC Monkey? A Rhesus monkey which is specially trained to quickly identify servers that need to be rebooted and get to them in a matter of seconds, regardless of datacenter location. After many years of studying datacenter operations, we found that the most time-consuming part of a server reboot was locating and getting to the server. Long rows of contiguous racks and servers nearly identical in appearance make it difficult for humans to easily navigate and find servers. However, Rhesus monkeys are much more agile and can cope with "forests" of racks, easily leaping over and swinging around rack poles to reach remote servers.
The monkeys are trained to locate servers in two stages. First, they are shown a map of the datacenter with the rack affected highlighted. Since the monkeys are normally above the racks, its very natural for them to utilize this "look down" of the datacenter to find the rack location. In the second stage, they are shown a representation of a rack with the relative locations of the servers in need of a reboot highlighted (picture attached). A fully-trained monkey can identify a rack and server location in under 2 seconds. After identification, it only takes about 5 seconds for the monkey to reach the rack and toggle the power switch on the proper server. This is about as much time as it takes for a browser page to reload .. nearly instant!
A datacenter with 5,000 servers can be handled by a team of 12-16 monkeys, even on a very busy day. They work on rotating shifts with plenty of resting time between reboot shifts. Rhesus monkeys are very mellow and not easily distracted, which makes them ideal for this sort of work. Each datacenter has a jungle atrium built-in to allow the monkeys some "time in the wild". Even though we have a flawless safety record to date, we have on-site zoologists and doctors just in case.
How cost-effective is this solution? Very! Even with a team of 12 monkeys in every datacenter, and the costs associated with caring for them, the cost per server is only $10 per month. This allows us to provide the technology to all of our customers at no cost and still maintain a healthy profit margin.