Re: who offers cheap (personal) 1U colo?
> $50/month at 40U rentable is $2000/rack/month if it's full. > after paying for 60A of power and 50Mbits/sec of transit > and whatever the rack rents for, the provider's gross margin > will be between 25% and 50%, out of which they have to pay > salaries. as a standalone business this makes no sense, but > at scale or as part of another business, $50/month @1U is > just about right. I've only seen a few comments on the business aspect of this, so I'd like to throw my two cents in. Given: at least certain Linux distributions are free to copy Given: the various BSD distributions are all free to copy Given: vmware workstation is a relatively low-cost product Given: Linux and BSD run in virtual machines on Vmware on Linux Question: Why can't a provider sell virtual PC colocation, instead of physical PC colocation? So instead of 40 physical machines per rack, why can't it be 80 or 160 or even more virtual machines, running on 40 physical Linux boxes? I think the economics could shift significantly under those circumstances. For personal colo the virtual CPU would probably be idle at least 99% of the time. My home servers usually are. Which means that when hosting 4 typical virtual machines a real CPU would still be mostly idling. Also a small IDE drive now is about 120 GB. Divide that by 4 and each colo still has 30 GB of disk space, more than enough for most needs. The hardware cost per "machine" certainly goes down, and other than the vmware licenses the OS software is "free", either BSD licensed free or GPL licensed "free". Either is good enough for this purpose. Is some hosting company already doing this?
Re: What's the best way to wiretap a network?
> You can plug a mini-hub in line and use that as a tap point to monitor > the stream. Up side is its cheap and easy. Down side is you have to > drop to half duplex. Not a problem in most situations but in some the > drop in performance can be an issue. Don't throw out your old hubs. It's hard to find a 10/100 hub for sale any more. Even the cheap consumer devices are switches. I picked up a $25 hub at Fry's a few weeks ago just in case I ever wanted to casually snoop some traffic. But Fry's is sold out. The netoptics.com link posted was "priceless". I'm always wary of simple products that are expensive enough to have a request for quote, rather than a price, on their web page.
Re: Is there a technical solution to spam?
> The solutions may well be found there but will be unimplementable > without much needed support from the operators - particularly the > major backbones - who currently turn a blind eye to protect their > revenue. Bingo. There's the crux of the problem. It needs to be elaborated on and emphasized, because most engineers have a blind spot about the business aspects of their industry (no matter what that industry is). There's a lot of wailing and gnashing of teeth, a lot of soul searching, a lot of angst here. All for naught. Many big network operators are selling bigger and bigger pipes to everyone so they can keep up with more and more spam. They make money on the increased traffic, even as they have these solemn terms and conditions in place about how they won't tolerate spam. The big network operators don't need to allow spammers to connect directly to their backbones. They make money by selling transit to other networks who sell transit to still other networks who then allow spammers to connect. Network operators are such a naive bunch of engineers. There's lots of money to be made just in transit for spam, and quite often the people who sign the paychecks for the engineers who post to this list are the very people who benefit. They understand this, why don't you? Every network operator should first try to get their own company to get serious about stopping spam. Top management has to be willing to do what it takes. E.g. de-peer, stop selling transit, etc. Until that happens the spam problem will keep getting worse. And if top management isn't interested, or won't agree to do anything meaningful, ask yourself why. And keep that in mind the next time you get paid.
Re: Spanning tree melt down ?
> Minimal social engineering plus a weak network security infrastructure > is a disaster waiting to happen for any major medical facility. You forgot to mention probable political infighting. And maybe inexperienced leadership. My favorite snippet from the article is: Dr. John Halamka, the former emergency-room physician who runs Beth Israel Deaconess Medical Center's gigantic computer network. Is a physician, after years of medical school, internship, residency, etc. the right person to be in charge of a "gigantic" computer network? Are arteries and veins the equivalent of fiber and CAT-5? I'd love to be the Cisco rep selling $3 million of new network equipment to this guy. What is the probability that he as ANY idea what "spanning tree protocol" means?
