Do you regularly consume your body weight in saccharine? No? Then I highly doubt the "pink stuff" will give you cancer. (Personally, I think the stuff tastes nasty.)
Unless you're smuggling shit by the gram, the lead sheilding will be huge and VERY heavy. That makes your common interstate car a non-target. Most mini-vans and station wagons would be dragging the ground with a ton of lead in there. A pick-up truck would be rather easy to visually inspect before pulling over. Large trucks and hummers, yeah, by all means, pull those a***oles over.:-)
If they have them that damned sensitive, why isn't the concrete walls in the subways setting it off? This just screams "alarmest" to me. The amount of material necessary to make a bomb, dirt or otherwise, is thousands of times greater than what they are setup to detect. In other words, if any actual bomb making kit passed within a mile of one of those sensors, it would almost certainly beep and then immediately catch fire. (think radar detector in a microwave oven.)
5 nines (99.999%) is 5min per year. 6 nines (99.9999%) is 30sec.
5 nines isn't always easy, but does exist. 6 nines also exists in the real world, despite the impossible requirement... think of fire supresion systems, lifesaving equipment, etc.; they work every time we need them, because people put a tremendous amount of work into keeping it that way.
Incorrect... 100% of my cable modem outages over the last year have been the result of TW breaking it. Software updates to various parts of the network -- which, as a network engineer, I know for a fact does not require turning off the customers for hours at a time.
The unexpected downtimes are generally rare events, even with an aging infrastructure.
I was born in '72. I can recall phone service being dead only twice. The last, a few years ago, was the result of the entire CO "crashing". (I don't know how a "crash" can cause a loss of loop current, but that was their story. Our DSLAM didn't lose power, so I don't know what they screwed up.) The other was a decade ago... trunk failure prevented calls out of the CO.
Bottom line, things that aren't supposed to happen, do sometimes happen.
power and telephone were life-and-death services
EXACTLY. That's the part people tend to gloss over. Seeing the latest Southpark episode is not a life or death situation. Likewise, your heart isn't going to explode because you cannot get to Yahoo! immediately.
Trusting one's life to a cell phone is a gamble. While they are a fairly stable technology, there are numerous troubling issues... Batteries don't last forever. Service isn't available everywhere. 911 calls aren't always routed to the most appropriate call center -- although it's much better than in years past. In an accident, your cell phone is just as likely to be damaged as you -- or worse, lost. etc.
You're missing the all-important DNS Voodoo of the Akamai system. When a user in ISP A's network looks up the host of your image tag, Akamai will return an IP within ISP A's network because they have a local Akamai cache server. Similarly, a user in ISP B's network will be directed to ISP B's local Akamai cache. For users in ISP C's network, who doesn't have a local cache, Akamai will return a cache server close to ISP C -- where "close" depends on network topology.
Everyone who hit your setup retreived the images from the same location.
I doubt a lot of their algorithms they use would pass the "obviousness" test...
They might not have been very obvious when Akamai started, but they're pretty obvious today. Crafting a DNS response based on who asked has been common practice for several years now -- "views" in bind. Akamai blends what with web proxies and network geography to send you to a proxy logically close to you, obviously preferring proxies within your ISP's network. ('tho not all ISP's host Akamai cache servers, which is sorta stupid as they'll give them to you. and they can save a huge amount of bandwidth.)
Yeah, they've gotten a little carried away with "requiring" service contracts. However, there are paths around it. I downloaded updated switch software (2950) last week without any contracts. Yes, there are little lock icons next to everything, but anyone with a CCO account (even a guest/non-contract user) can download them. Getting past the front door is a bit of a task as you cannot simply head to the software center like you used to.
If you email tac, they'll send you whatever software you're entitled to.
"no additional charge"... because they stuck it into the purchase price. The bare metal is just as cheap as everything else -- some times surprisingly cheap despite the $$$$ price tag.
Nobody. Cisco list prices are printing catalogs. Even Cisco doesn't sell stuff at list price.
(And yes, there are 100% couterfit hardware. Last I checked, cisco doesn't have any manufacturing centers in China. Everything I've ever seen was made in either Mexico or Malaysia.)
IN SOFTWARE. That's generally ok for linux machines. However, that's not ok for "real routers" as they do almost everything with dedicated hardware; they have very small, cheap, underpowered processors for dealing with thousands of packets per second. Go build yourself a linux router out of an old 386 and see how well it deals with IPv6 -- I can tell you from experience, the answer is "poorly." IPv6 means new switching hardware, substantially faster processors, and significantly more memory. Just because I can load IPv6 aware software on a Cisco 1601 doesn't mean it's a good plan to do so.
There are no 10 year old backbone routers still in service on any backbone. Anywhere.
ABSOULTELY, 100%, wrong. You've obviously never worked for an ISP. They do not go around replacing their entire infrastructure every few years. The Teir-1 providers won't even think about installing anything until it's been throughly tested -- a process that can take years. For example, you won't find hardware Cisco just started manufacturing on UUNet's backbone.
SSL, SSH, VPNs, and that "plethora of other protocols" are all very easily identifiable as not p2p traffic. SSL is immediately identifiable as SSL; even if I cannot decode it, I know it's SSL and can even know what server they're asking to talk to. Plus, it's pretty easy to tell the destination is a bank. Bittorrent doesn't use SSL -- at least not between peers; many clients support https tracker urls. SSH is, again, immediately identifiable. VPN traffic doesn't look anything even remotely like p2p traffic -- or do you usually open and close vpn connections to dozens of systems per minute?
I hate to break it to the nuts at Bittorent, Inc., but this isn't going to work AT ALL. a) All it takes it ONE (1) client not scrambing communications to immediately compromise the system. Once an info_hash is known, your only secret is no longer secret. b) The proposed caching and reuse of encrypted data opens the door to near realtime attack. Client behavior is very predictable... after contacting a tracker -- which, in the absence of SSL, is trivial to spot -- there's a spike of outbound connections. You only have to get one ip+port pair correct to recover the 160bit key and thus the entire peer list.
In today's world, we have these things called "line powered remote DSLAMs". They aren't very expensive for any real ISP, but it does take some work to hang the thing from a poll and wire it into the network. Bellsouth apparently has one hanging on a poll near my parent's house out in the sticks -- doesn't offer the "xtreme 6.0", but DSL is available. I don't know if this was before or after TW got their head out of their a** and extended cable the 1 mile necessary to reach everyone. (which they'd refused to do for 30 years.)
By virtue pf them never having had the right to make and sell you the device, yes, you can be bared from using it and/or have it taken away (presumablly with a refund of your purchase price.)
Ask the millions of people using MySQL. Most of the time, it's unnecessary -- think of what most people use a database for... blogs, torrent sites, catalogs, wiki's, etc. The rest of the time, your hardware and software are generally stable enough that it's often an unnecessary performance drain. And in today's world, it takes very specialized setups to be 100% certain data has actually been commited to disk. LJ learned that lesson the hard way... Has the hard drive actually written the data it says it has?
Postgres has had the "no fsync" option for over a decade. This is no different.
If it was tons of power, how could they make 802.11n adapters for laptops
Simple... laptops don't run their wireless adapters anywhere near max-power. And they don't run them continuously. (unless you've intentionally disabled power management.)
FPGA != ASIC I've taken a lot of cisco gear apart; very few have had real, custom logic in them. (and that was very expensive gear.) Bugs in ASICs are very difficult and expensive to fix. FPGAs are simple to reprogram; that's why they show up in everything.
Encryption is now done in hardware even in the cheapest g/n chipsets
Negative. WEP is done in hardware. WPA/WPA2 are done in software in every router I've looked at -- based on broadcom, atheros, ralink, and intel chips. Most of the broadcom cpus have a crypto processor in them, however no linux based firmware uses it. (the vxworks based ones might, but I've disassembled them to see.)
Exactly. The Linksys WRT300N sitting on my desk has a 12V 1000mA power brick. That's a MAX power draw of 12W. Usually it's not even that. The 600N dual radio, gigabit beast has a 2A brick -- given it's space heating abilities, it's using most of it.
Slashdot Mirror
"... collapse the solar system into an ultra dense particle about the size of a pea." -- Dr. Longbore.
Do you regularly consume your body weight in saccharine? No? Then I highly doubt the "pink stuff" will give you cancer. (Personally, I think the stuff tastes nasty.)
Unless you're smuggling shit by the gram, the lead sheilding will be huge and VERY heavy. That makes your common interstate car a non-target. Most mini-vans and station wagons would be dragging the ground with a ton of lead in there. A pick-up truck would be rather easy to visually inspect before pulling over. Large trucks and hummers, yeah, by all means, pull those a***oles over. :-)
... or a doctor's note.
If they have them that damned sensitive, why isn't the concrete walls in the subways setting it off? This just screams "alarmest" to me. The amount of material necessary to make a bomb, dirt or otherwise, is thousands of times greater than what they are setup to detect. In other words, if any actual bomb making kit passed within a mile of one of those sensors, it would almost certainly beep and then immediately catch fire. (think radar detector in a microwave oven.)
5 nines (99.999%) is 5min per year. 6 nines (99.9999%) is 30sec.
5 nines isn't always easy, but does exist. 6 nines also exists in the real world, despite the impossible requirement... think of fire supresion systems, lifesaving equipment, etc.; they work every time we need them, because people put a tremendous amount of work into keeping it that way.
Incorrect... 100% of my cable modem outages over the last year have been the result of TW breaking it. Software updates to various parts of the network -- which, as a network engineer, I know for a fact does not require turning off the customers for hours at a time.
The unexpected downtimes are generally rare events, even with an aging infrastructure.
Bottom line, things that aren't supposed to happen, do sometimes happen.EXACTLY. That's the part people tend to gloss over. Seeing the latest Southpark episode is not a life or death situation. Likewise, your heart isn't going to explode because you cannot get to Yahoo! immediately.
Trusting one's life to a cell phone is a gamble. While they are a fairly stable technology, there are numerous troubling issues... Batteries don't last forever. Service isn't available everywhere. 911 calls aren't always routed to the most appropriate call center -- although it's much better than in years past. In an accident, your cell phone is just as likely to be damaged as you -- or worse, lost. etc.
You're missing the all-important DNS Voodoo of the Akamai system. When a user in ISP A's network looks up the host of your image tag, Akamai will return an IP within ISP A's network because they have a local Akamai cache server. Similarly, a user in ISP B's network will be directed to ISP B's local Akamai cache. For users in ISP C's network, who doesn't have a local cache, Akamai will return a cache server close to ISP C -- where "close" depends on network topology.
Everyone who hit your setup retreived the images from the same location.
Yeah, they've gotten a little carried away with "requiring" service contracts. However, there are paths around it. I downloaded updated switch software (2950) last week without any contracts. Yes, there are little lock icons next to everything, but anyone with a CCO account (even a guest/non-contract user) can download them. Getting past the front door is a bit of a task as you cannot simply head to the software center like you used to.
If you email tac, they'll send you whatever software you're entitled to.
"no additional charge"... because they stuck it into the purchase price. The bare metal is just as cheap as everything else -- some times surprisingly cheap despite the $$$$ price tag.
Nobody. Cisco list prices are printing catalogs. Even Cisco doesn't sell stuff at list price.
(And yes, there are 100% couterfit hardware. Last I checked, cisco doesn't have any manufacturing centers in China. Everything I've ever seen was made in either Mexico or Malaysia.)
SSL, SSH, VPNs, and that "plethora of other protocols" are all very easily identifiable as not p2p traffic. SSL is immediately identifiable as SSL; even if I cannot decode it, I know it's SSL and can even know what server they're asking to talk to. Plus, it's pretty easy to tell the destination is a bank. Bittorrent doesn't use SSL -- at least not between peers; many clients support https tracker urls. SSH is, again, immediately identifiable. VPN traffic doesn't look anything even remotely like p2p traffic -- or do you usually open and close vpn connections to dozens of systems per minute?
I hate to break it to the nuts at Bittorent, Inc., but this isn't going to work AT ALL. a) All it takes it ONE (1) client not scrambing communications to immediately compromise the system. Once an info_hash is known, your only secret is no longer secret. b) The proposed caching and reuse of encrypted data opens the door to near realtime attack. Client behavior is very predictable... after contacting a tracker -- which, in the absence of SSL, is trivial to spot -- there's a spike of outbound connections. You only have to get one ip+port pair correct to recover the 160bit key and thus the entire peer list.
Tell it to LiveJournal, who had the expensive hardware SCSI RAID, and it STILL DIDN'T turn off the drive's write cache.
In today's world, we have these things called "line powered remote DSLAMs". They aren't very expensive for any real ISP, but it does take some work to hang the thing from a poll and wire it into the network. Bellsouth apparently has one hanging on a poll near my parent's house out in the sticks -- doesn't offer the "xtreme 6.0", but DSL is available. I don't know if this was before or after TW got their head out of their a** and extended cable the 1 mile necessary to reach everyone. (which they'd refused to do for 30 years.)
Actually, there are reports of them ripping the copper up. (just, fyi)
By virtue pf them never having had the right to make and sell you the device, yes, you can be bared from using it and/or have it taken away (presumablly with a refund of your purchase price.)
Postgres has had the "no fsync" option for over a decade. This is no different.
FPGA != ASIC I've taken a lot of cisco gear apart; very few have had real, custom logic in them. (and that was very expensive gear.) Bugs in ASICs are very difficult and expensive to fix. FPGAs are simple to reprogram; that's why they show up in everything.
Exactly. The Linksys WRT300N sitting on my desk has a 12V 1000mA power brick. That's a MAX power draw of 12W. Usually it's not even that. The 600N dual radio, gigabit beast has a 2A brick -- given it's space heating abilities, it's using most of it.