If your enterprise is such that you have a few hundred routers, then I'd certainly hope that you'd have ponied up for Cisco Works, which would then allow you to push out the upgrades in an automated manner.
Of course, there are also freely available perl and expect scripts out there that would allow you to do the same thing.
Most Cisco code updates do not require TAC intervention, or email swapping. This is an isolated case.
Also, I haven't had to mail TAC yet for any of the routers (30, and counting) I've had to upgrade. My new code has been available throught the traditional channel (Cisco's Software Center).
People that are having to mail the TAC are doing so because they have no support contract (thus, no access to the Cisco Software Center), or because the code for their specific platform doesn't appear to be available through the Software center.
I've always held that IE's stranglehold on the web, and nose-thumbing at W3C standards, would come to a rapid close the day that AOL replaced IE with Mozilla (Netscape, Firebird, whatever) in the latest and greatest AOL client software update.
30 Million users is no joke... When an Internet commerce site starts getting complaints that AOL users can't navigate their site, or buy their wares, they're going to clean up that spaghetti web code quickly.
We tend to dismiss AOL users, and their chosen ISP, as being the Internet's tricycle set. Let's not overlook their potential as an economic engine for change.
If you only play "the nuts," then you'll never get anywhere.
The other players will know that if you stay in the hand, it's because you've got something strong, and they'll fold (Especially the tight players). You'll never get them to commit their stacks... You won't be able to trap...
You have to play some meek hands, steal some pots on bluffs, and periodically, you've got to show them that you beat them out of a pot on the back of a 4-7 off-suit. That way, when you get dealt the monster, they won't be able to tell whether or not you've really got the goods.
Thanks for the tip... I played there for quite a while last night, and their Java client is much better than a lot of the full blown Windows clients I've seen.
Card counting is a Blackjack thing... Not a poker thing.
You can't count cards in Poker. There's a single deck, and a fresh shuffle every hand. Additionally, you don't get to see the cards of the players that end up folding. There's simply no way to determine if that card you need is still in the deck, or if one of the other players has it in their hand, or if it's already been folded.
In Blackjack, you get to see every card, so you can keep a running tally of how many facecards have come and gone between shuffles. Most Casinos, however, utilize 6 decks of cards at their Blackjack tables, so the fact that an Ace has come and gone doesn't mean what it would, advantagewise, in a single deck game.
Professional Blackjack players count cards thusly (and this is rough, as I'm not a card counter, nor a professional Blackjack player):
High Cards left in the deck favor the Player (Bettering the chances of being dealt a good hand, or building a good hand on a hit, and bettering the chances that the House will bust on a forced hit from a hand with a value of less than 16)
Low Cards left in the deck favor the House (Reducing the odds of the House busting on a forced hit with a hand less than 16)
A Player keeps a running Tally of what has been dealt
Every High Card that goes by is assigned a negative integer value
Every Low Card that goes by is assigned a positive integer value
Toward the end of the shuffle, if the running Tally is in Positive Territory, the Player should be betting aggressively (The greater the tally, the more aggressively the bets should be)
Toward the end of the shuffle, if the running Tally is in Negative Territory, the Player should be betting conservatively (The lower the tally, the more conservative the bets should be)
It's all predicated on the ability to determine when the conditions favor the Player, and when the conditions favor the House. You maximize your betting when the conditions are favorable for you. You minimize your betting when the conditions are favorable to the House. It's a very effective method for finding the sweet spot on the risk v. reward curve, and managing your stack.
In spite of what you may have heard, particularly from the Casinos themselves, there is nothing illegal about counting cards in Blackjack. You cannot be arrested. You cannot be fined. It's perfectly legal.
However, just like any other business in the world, the Casino reserves the right to refuse anyone's business for any reason. If they suspect you are a card counter, a ringer, they can and will ask you to leave. You'll go into their database, and the Casino's are sharing more and more of their information between each other. You will end up Blackballed.
If you're going to Vegas to play Blackjack, you need to be playing it at Binion's Horseshoe... They have more Single Deck Blackjack games than any other Casino in Vegas.
Moneymaker won the 2003 tourney just this past May (And is thus the reigning Champion).
The 2002 winner (Robert V., from Brooklyn, NY) was put out in the early rounds of the 2003 tourney, not that it's important...
There are plenty of tournament winners that get their asses kicked on the very next event. Play style, skill, and betting strategies are all part of the equation, but at some point, no matter how good you are, eventually everyone runs into a tough beat because sometimes the other player simply has good cards.
In Limit Hold-em, the better players should always end up winning... Limit is lacking the "all in" feature of the No Limit game. In No Limit, a fluke can put the better player out of the game.
Gus Hansen is a wildman. I've been watching the World Poker Tour all season, and while there are no shortage of aggressive players, I haven't seen anybody bully a table like Hansen does.
Redundant perhaps, but not Flamebait! When wiggys posted his comment mere seconds before I did, he got modded as Funny.
Besides, he didn't even fully commit to the bash... It was very timid, put into the form of a question even. I went "all in" on the jab.
Ahh... Saner moderators are checking in now, thank goodness. Well done gentlemen, well done. I fully expect Budweiser to run a "Real Men of Genius" radio ad dedicated to Slashdot moderators Real Soon Now[tm]. (Mr. Slashdot comment moderator... You know the tune.)
While I'm glad to see someone mentioning OSPF (Which assigns a numerical value to each link within an area based on the available bandwidth, and determines best path by applying the Djikstra algorithm... This is also how Spanning-tree Protocol (802.1d) functions.), it's not how traffic is routed on "The Internet."
The Internet uses Border Gateway Protocol (BGPv4), which is a Path-Vector routing protocol. OSPF is a Link-State routing protocol.
OSPF is considered an IGP (Interior Gateway Protocol). It can be used within an autonomous system, but has no place in determining path selection for traffic between automonous systems.
As far as IGPs go, there's only one that would be capable of handling the routing table for the entire Internet, and it's not OSPF. IS-IS, Intermediate Syetem to Intermediate System, is another Link-State protocol developed by the OSI during the same period when OSPF was being developed.
They share a lot of similar features, and address all of the same shortcomings inherent to Distance-Vector routing protocols (RIP, IGRP). You can actually redistribute the full Internet routing table from BGP into IS-IS, and it will handle the strain.
Aside from the ability to handle astronomically large routing tables, IS-IS has one additional feature that sets it above OSPF: No requirement for a single backbone area (Area 0, in OSPF speak).
OSPF is not particularly well suited to "meshy" environments, due to the need for a single, clearly defined backbone area (In OSPF, all traffic between non-backbone areas must traverse Area 0). IS-IS alleviates this requirement. There can be multiple Inter-area paths, which can be very useful in a complex network.
Of course, the pool of IS-IS savvy network engineers is far smaller than that of the OSPF disciples, so you don't see it in use very often. The exception is in the service provider space. Big ISPs, and Backbone Carriers frequently utilize IS-IS when an IGP is called for, notably for it's ability to handle large routing tables.
(Don't get me wrong... I'm a fan of OSPF, but much like the programmer folks like to say, "It's just a tool in the toolbox." The savvy network engineer will utilize the Routing Protocol which best suits his requirements (In some cases he'll use more than one), just as the savvy programmer with utilize the programming language that best suits his requirements.)
In summary:
Djikstra Good.
Dijkstra Algorithm Good.
OSPF Good.
OSPF != Internet Routing.
BGPv4 = Internet Routing.
IS-IS Good.
IS-IS > OSPF.
Routing Protocol Selection every bit as important as Programming Language Selection.
For instance, there should definately have been a crack about French authorities busting heads over the fact that this streaming Japanese content wasn't properly voiced-over in French, with Japanese subtitles... Preserving their culture, and all that.
**chuckle** French Authorities Busting Heads! Sometimes, I kill myself.
I've never had any personal interaction with Mr. Becker, so I can't really comment on his attitude, but I have been running various flavors of Linux for the past 8 years, all of which have had Ethernet cards of one sort or another.
Far more often than not, they've performed flawlessly, and I have Don Becker to thank for it. If you use Ethernet on Linux, you're either using a driver developed by Mr. Becker almost entirely on his own, or you're using a driver kludged together by someone else that is almost entirely derived from Mr. Becker's code.
I didn't even realize he had any affiliation with the Beowulf project until I read this story... I'd only known of Becker as Mr. Ethernet on Linux.
I was so excited to be submitting a story involving a Beowulf cluster, that I didn't think the concluding statement through...
It should have read:
Pii writes "News.com is running a story about Penguin Computing acquiring Scyld Computing, a company founded by Donald Becker, of linux ethernet driver and Beowulf cluster fame. Becker will stay on as Penguin's Chief Technology Officer, and the companies claim they don't expect any layoffs as a result of the merger. A naked and petrified Natalie Portman was otherwise occupied, and couldn't be reached for comment. Hot Grits, Hot Grits, Hot Grits!"
A pox on me for not having had the discipline to execute this story correctly.
The point was not that a solution for a 30 Campus WAN would directly translate to a solution for a multi-story, multi-family dwelling.
The point was that once you obtain the space in the co-lo, and the bandwidth, you can hang a lot of network behind it pretty inexpensively.
The reason all of the co-los sit empty is two-fold:
There are a billion co-lo facilities that have been built, many of which had broken ground prior to the bursting of the bubble, and there's no longer the kind of demand that was anticipated 4 years ago, and
You idiot ISPs have done a piss poor job at articulating the advantages of co-location, and have failed to instill enough fear in your potential customers that they're ill-prepared to handle power, or circuit disruptions, and that every time they can't receive email, it's costing them money.
While you continue to try to keep up, you've completely missed what I've outlined above, so let me break it down for you:
In addition to presenting a pretty good solution for the "Ask Slashdot" topic at hand, I've shown how that same solution can be applied to multiple additional locations, and scaled it up to a working business model.
Keep arguing with me about T-1 tarrifs.
And for my final throw away smart-assed remark, you claim to work for a Tier-2 ISP.
If everyone at your company has the same winning personality, and tunnel-vision, I have to conclude that "Tier-2", and "2nd Rate" are interchangable terms.
Thank you, and Goodnight. Remember to tip your waitresses...
"Kid" was probably unnecessary... My apologies. I've stirred up a flamewar of USENET proportions, which is fun, but there are so many "Anonymous Cowards" involved (At least, I think there's more than one), it's hard to keep them straight. I was exceedingly harsh, but let's be honest... All you anonymous cowards look alike.
I too like the 4000 chassis, and I don't mean to diminish it's capablilities. It's kick-ass, but expensive, and for this scenario, I think it ill suited because:
The solution doesn't require significant port density in any single location (In a multistory facility, best practice means connecting access-layer devices in the horizontal plane, not the vertical).
If port density were an issue at the core, a 48 port Cat 3550 would fit the bill pretty well, still be far less expensive than a Cat 4000.
The Cat 3550 provides all of the Layer-3 functionality necessary to act as the core layer device.
The Cat 3550 is vastly less expensive than even a minimally configured Cat 4000 providing the same features (The key feature being the Layer-3 capability as the Layer-3 enabled Supervisor will run in excess of $15,000, and that doesn't include the chassis, or the 48 port module).
This application, as described, is not about providing a network drop for hundreds of users in an office cube farm. It's about providing one (or maybe two) drops per user dwelling, in a multistory apartment/condo building.
As for the Cat 2900 series being discontinued, you're right... I don't sell them, I just integrate them, and I'm still deploying them while supplies last. I suppose they're pushing Cat 3500s at the access-layer platforms now, right?
Hey BONER, that's an inbound access-list placed on the individual condo vlan interface, and has no affect on return traffic...
Why is it that every Anonymous Coward thinks they know some shit? If you're gonna call people names, why don't you stand up while doing it? SlashdotIDs are free.
You haven't even provided a name, much less any kind of credential, testifying to your level of expertise.
You have a cursory understanding of the OSI model, which is great in preparing for th Net+ exam, I suppose, but really doesn't qualify you to sift through postings on Slashdot, and determine who does, and does not, know what they are talking about.
I've been working exclusively in the network space for the past 13 years, primarily with Cisco routers and switches for the past 9. I've built networks for Fortune 10 companies, and Government agencies.
When your bank uploads its financial transaction information to the Federal Reserve every night, that data goes across a network that I built.
When you filled out your application for a US Passport, a scanned image of that document traversed a network at the State Department that I built.
When a Telco switch in Saudi Arabia experiences some kind of difficulty, error messages and telemetry data ride back to the central montitoring center in Riyadh across a network that I built.
When government offices in remote villages in Afghanistan place their VoIP or Video conferenceing calls back to Kabul, that data rides over a network that I built.
Ever build a campus network for a 70 building, 10,000 user Marine Base? I have.
Ever build a network for a 30 Campus University? I have.
Ever configured a 1500+ site frame relay network? I have.
Tell me something about BGP community settings, or Multi-Exit Descriminators? Tell me something about IS-IS, the only interior routing protocol that is capable of handling the entire BGP table when redistributed into it. Tell me about Type-7 LSAs, or the trouble with OSPF in NMBA environments. Tell me about tuning Spanning-tree in my campus. Tell me about NBAR, CBAC, and Reflexive access-lists.
Last but not least... Please, Sir, tell me more about how there's no such thing as a Layer-3 switch... I so much want to learn.
Look, if you'd like to just whip out our dicks, and measure, great...
Yes, I'm providing ballpark pricing. I'm not going to get a valid price quote for a discussion on Slashdot.
I'm no stranger to Telco. I currently work for one of the largest Verizon authorized agents in the Mid-Atlantic. We are also authorized agents for Qwest, and AT&T. Prior to working here, I worked for Sprint, and before that, AT&T.
Just last Summer, I put in a WAN for a 30 campus University, with the head-end at an AT&T co-lo in Northern Virginia, and Internet access for the whole network is provided via the co-lo. My point is this: I know of which I speak. Just because you haven't done something, don't assume that it can't be done.
Layer-3 switching, Anonymous Jackass, is a switch that is also capable of making forwarding decisions based on Layer-3 (Network) information.
Examples of Layer-3 switches are numerous, but since I'm a Cisco guy, allow me to direct you to:
Cisco Catalyst 3550 Series (w/either the SMI or EMI Images)
Cisco Catalyst 4000 Series (w/the Supervisor 3 or Supervisor 4)
Cisco Catalyst 5000 Series (w/the Route Switch Module)
Cisco Catalyst 6000 Series (w/the Multi-layer Switching Module, or MSFC I or II (Multi-layer Switching Feature Card))
There are plenty of others in Cisco's lineup, and from a wide range of other manufactures.
The key is that any of these platforms can perform inter-Vlan routing without the need of an external router. The routing feature has been integrated into the switch. (And is typically performed in hardware, through the use of ASICs)
Slashdot Mirror
You have a Cisco 2610...
What Feature pack?
- ENTERPRISE PLUS
- ENTERPRISE PLUS IPSEC 3DES
- ENTERPRISE PLUS IPSEC 56
- ENTERPRISE/FW/IDS PLUS IPSEC 3DES
- ENTERPRISE/FW/IDS PLUS IPSEC 56
- ENTERPRISE/SNASW PLUS
- ENTERPRISE/SNASW PLUS IPSEC 3DES
- ENTERPRISE/SNASW PLUS IPSEC 56
- IP
- IP PLUS
- IP PLUS IPSEC 3DES
- IP PLUS IPSEC 56
- IP/FW/IDS
- IP/FW/IDS PLUS IPSEC 3DES
- IP/FW/IDS PLUS IPSEC 56
- IP/H323
- IP/IPX/AT/DEC
- IP/IPX/AT/DEC PLUS
- IP/IPX/AT/DEC/FW/IDS PLUS
- REMOTE ACCESS SERVER
That's just the available images for the 2610, 12.1(20)...Of course, there are also freely available perl and expect scripts out there that would allow you to do the same thing.
Also, I haven't had to mail TAC yet for any of the routers (30, and counting) I've had to upgrade. My new code has been available throught the traditional channel (Cisco's Software Center).
People that are having to mail the TAC are doing so because they have no support contract (thus, no access to the Cisco Software Center), or because the code for their specific platform doesn't appear to be available through the Software center.
I have to ask... What exactly were you promised, and how does radio fail to live up to it?
I listen to talk radio 3+ hours a day while communting, and I'd estimate it's value to be far higher than what I've paid for it.
30 Million users is no joke... When an Internet commerce site starts getting complaints that AOL users can't navigate their site, or buy their wares, they're going to clean up that spaghetti web code quickly.
We tend to dismiss AOL users, and their chosen ISP, as being the Internet's tricycle set. Let's not overlook their potential as an economic engine for change.
It's a facintating and fun game, and if a person was skilled enough to make a living at it, I can't imagine going back to a regular career.
The other players will know that if you stay in the hand, it's because you've got something strong, and they'll fold (Especially the tight players). You'll never get them to commit their stacks... You won't be able to trap...
You have to play some meek hands, steal some pots on bluffs, and periodically, you've got to show them that you beat them out of a pot on the back of a 4-7 off-suit. That way, when you get dealt the monster, they won't be able to tell whether or not you've really got the goods.
Thanks for the tip... I played there for quite a while last night, and their Java client is much better than a lot of the full blown Windows clients I've seen.
You can't count cards in Poker. There's a single deck, and a fresh shuffle every hand. Additionally, you don't get to see the cards of the players that end up folding. There's simply no way to determine if that card you need is still in the deck, or if one of the other players has it in their hand, or if it's already been folded.
In Blackjack, you get to see every card, so you can keep a running tally of how many facecards have come and gone between shuffles. Most Casinos, however, utilize 6 decks of cards at their Blackjack tables, so the fact that an Ace has come and gone doesn't mean what it would, advantagewise, in a single deck game.
Professional Blackjack players count cards thusly (and this is rough, as I'm not a card counter, nor a professional Blackjack player):
It's all predicated on the ability to determine when the conditions favor the Player, and when the conditions favor the House. You maximize your betting when the conditions are favorable for you. You minimize your betting when the conditions are favorable to the House. It's a very effective method for finding the sweet spot on the risk v. reward curve, and managing your stack.
In spite of what you may have heard, particularly from the Casinos themselves, there is nothing illegal about counting cards in Blackjack. You cannot be arrested. You cannot be fined. It's perfectly legal.
However, just like any other business in the world, the Casino reserves the right to refuse anyone's business for any reason. If they suspect you are a card counter, a ringer, they can and will ask you to leave. You'll go into their database, and the Casino's are sharing more and more of their information between each other. You will end up Blackballed.
If you're going to Vegas to play Blackjack, you need to be playing it at Binion's Horseshoe... They have more Single Deck Blackjack games than any other Casino in Vegas.
Moneymaker won the 2003 tourney just this past May (And is thus the reigning Champion).
The 2002 winner (Robert V., from Brooklyn, NY) was put out in the early rounds of the 2003 tourney, not that it's important...
There are plenty of tournament winners that get their asses kicked on the very next event. Play style, skill, and betting strategies are all part of the equation, but at some point, no matter how good you are, eventually everyone runs into a tough beat because sometimes the other player simply has good cards.
In Limit Hold-em, the better players should always end up winning... Limit is lacking the "all in" feature of the No Limit game. In No Limit, a fluke can put the better player out of the game.
Gus Hansen is a wildman. I've been watching the World Poker Tour all season, and while there are no shortage of aggressive players, I haven't seen anybody bully a table like Hansen does.
Redundant perhaps, but not Flamebait! When wiggys posted his comment mere seconds before I did, he got modded as Funny.
Besides, he didn't even fully commit to the bash... It was very timid, put into the form of a question even. I went "all in" on the jab.
Ahh... Saner moderators are checking in now, thank goodness. Well done gentlemen, well done. I fully expect Budweiser to run a "Real Men of Genius" radio ad dedicated to Slashdot moderators Real Soon Now[tm]. (Mr. Slashdot comment moderator... You know the tune.)
Sorry... Just a rittle Japan bashing humor. Move along.
The Internet uses Border Gateway Protocol (BGPv4), which is a Path-Vector routing protocol. OSPF is a Link-State routing protocol.
OSPF is considered an IGP (Interior Gateway Protocol). It can be used within an autonomous system, but has no place in determining path selection for traffic between automonous systems.
As far as IGPs go, there's only one that would be capable of handling the routing table for the entire Internet, and it's not OSPF. IS-IS, Intermediate Syetem to Intermediate System, is another Link-State protocol developed by the OSI during the same period when OSPF was being developed.
They share a lot of similar features, and address all of the same shortcomings inherent to Distance-Vector routing protocols (RIP, IGRP). You can actually redistribute the full Internet routing table from BGP into IS-IS, and it will handle the strain.
Aside from the ability to handle astronomically large routing tables, IS-IS has one additional feature that sets it above OSPF: No requirement for a single backbone area (Area 0, in OSPF speak).
OSPF is not particularly well suited to "meshy" environments, due to the need for a single, clearly defined backbone area (In OSPF, all traffic between non-backbone areas must traverse Area 0). IS-IS alleviates this requirement. There can be multiple Inter-area paths, which can be very useful in a complex network.
Of course, the pool of IS-IS savvy network engineers is far smaller than that of the OSPF disciples, so you don't see it in use very often. The exception is in the service provider space. Big ISPs, and Backbone Carriers frequently utilize IS-IS when an IGP is called for, notably for it's ability to handle large routing tables.
(Don't get me wrong... I'm a fan of OSPF, but much like the programmer folks like to say, "It's just a tool in the toolbox." The savvy network engineer will utilize the Routing Protocol which best suits his requirements (In some cases he'll use more than one), just as the savvy programmer with utilize the programming language that best suits his requirements.)
In summary:
So many possibilities squandered.
For instance, there should definately have been a crack about French authorities busting heads over the fact that this streaming Japanese content wasn't properly voiced-over in French, with Japanese subtitles... Preserving their culture, and all that.
**chuckle** French Authorities Busting Heads! Sometimes, I kill myself.
Far more often than not, they've performed flawlessly, and I have Don Becker to thank for it. If you use Ethernet on Linux, you're either using a driver developed by Mr. Becker almost entirely on his own, or you're using a driver kludged together by someone else that is almost entirely derived from Mr. Becker's code.
I didn't even realize he had any affiliation with the Beowulf project until I read this story... I'd only known of Becker as Mr. Ethernet on Linux.
I was so excited to be submitting a story involving a Beowulf cluster, that I didn't think the concluding statement through...
It should have read:
A pox on me for not having had the discipline to execute this story correctly.And for your list...
9. VD (The creeping crud)
The point was not that a solution for a 30 Campus WAN would directly translate to a solution for a multi-story, multi-family dwelling.
The point was that once you obtain the space in the co-lo, and the bandwidth, you can hang a lot of network behind it pretty inexpensively.
The reason all of the co-los sit empty is two-fold:
While you continue to try to keep up, you've completely missed what I've outlined above, so let me break it down for you:
In addition to presenting a pretty good solution for the "Ask Slashdot" topic at hand, I've shown how that same solution can be applied to multiple additional locations, and scaled it up to a working business model.
Keep arguing with me about T-1 tarrifs.
And for my final throw away smart-assed remark, you claim to work for a Tier-2 ISP.
If everyone at your company has the same winning personality, and tunnel-vision, I have to conclude that "Tier-2", and "2nd Rate" are interchangable terms.
Thank you, and Goodnight. Remember to tip your waitresses...
I too like the 4000 chassis, and I don't mean to diminish it's capablilities. It's kick-ass, but expensive, and for this scenario, I think it ill suited because:
This application, as described, is not about providing a network drop for hundreds of users in an office cube farm. It's about providing one (or maybe two) drops per user dwelling, in a multistory apartment/condo building.
As for the Cat 2900 series being discontinued, you're right... I don't sell them, I just integrate them, and I'm still deploying them while supplies last. I suppose they're pushing Cat 3500s at the access-layer platforms now, right?
Why is it that every Anonymous Coward thinks they know some shit? If you're gonna call people names, why don't you stand up while doing it? SlashdotIDs are free.
You haven't even provided a name, much less any kind of credential, testifying to your level of expertise.
You have a cursory understanding of the OSI model, which is great in preparing for th Net+ exam, I suppose, but really doesn't qualify you to sift through postings on Slashdot, and determine who does, and does not, know what they are talking about.
I've been working exclusively in the network space for the past 13 years, primarily with Cisco routers and switches for the past 9. I've built networks for Fortune 10 companies, and Government agencies.
Ever build a campus network for a 70 building, 10,000 user Marine Base? I have.
Ever build a network for a 30 Campus University? I have.
Ever configured a 1500+ site frame relay network? I have.
Tell me something about BGP community settings, or Multi-Exit Descriminators? Tell me something about IS-IS, the only interior routing protocol that is capable of handling the entire BGP table when redistributed into it. Tell me about Type-7 LSAs, or the trouble with OSPF in NMBA environments. Tell me about tuning Spanning-tree in my campus. Tell me about NBAR, CBAC, and Reflexive access-lists.
Last but not least... Please, Sir, tell me more about how there's no such thing as a Layer-3 switch... I so much want to learn.
Yes, I'm providing ballpark pricing. I'm not going to get a valid price quote for a discussion on Slashdot.
I'm no stranger to Telco. I currently work for one of the largest Verizon authorized agents in the Mid-Atlantic. We are also authorized agents for Qwest, and AT&T. Prior to working here, I worked for Sprint, and before that, AT&T.
Just last Summer, I put in a WAN for a 30 campus University, with the head-end at an AT&T co-lo in Northern Virginia, and Internet access for the whole network is provided via the co-lo. My point is this: I know of which I speak. Just because you haven't done something, don't assume that it can't be done.
Layer-3 switching, Anonymous Jackass, is a switch that is also capable of making forwarding decisions based on Layer-3 (Network) information.
Examples of Layer-3 switches are numerous, but since I'm a Cisco guy, allow me to direct you to:
- Cisco Catalyst 3550 Series (w/either the SMI or EMI Images)
- Cisco Catalyst 4000 Series (w/the Supervisor 3 or Supervisor 4)
- Cisco Catalyst 5000 Series (w/the Route Switch Module)
- Cisco Catalyst 6000 Series (w/the Multi-layer Switching Module, or MSFC I or II (Multi-layer Switching Feature Card))
There are plenty of others in Cisco's lineup, and from a wide range of other manufactures.The key is that any of these platforms can perform inter-Vlan routing without the need of an external router. The routing feature has been integrated into the switch. (And is typically performed in hardware, through the use of ASICs)
Get yourself a SlashdotID... They're free.
It's not uncommon for people living in a condo or apartment building to not even have met their neighbors (Particularly in the Northeast:) ).
I suspect that for most people, it's simply about obtaining Internet access, and by taking care of economies of scale, everyone can benefit.