So what were their titles, and where can I get a copy of the content?

Digital Watches, which are a pretty neat idea, were using ASICs in the 70's.

Yes; however, these were a different kind of ASIC. The kind of ASIC you would find in a watch is not performing a complex computational task by current standards.

You will also find certain ASICs in modern computers and servers which are part of the chipset.

When we are concerned about the difference between an ASIC-based forwarding architecture; we are primarily interested in tasks which cannot be executed efficiently at the same capacity on a general purpose FPGA or microprocessor, not the number of units which are being manufactured.

It turns out there are multiple different reasons ASICs become more cost-effective; when you would need a ton more general purpose processor wiring and power consumption to accomplish the same job that a well-designed application-specific processor constructed from ASICs could be more well-suited than a general purpose system using a generic computer board to complete the task, or a bunch of generic FPGAs and ICs glued together and programmed to do it in software.

considering they can easily take up less than a tenth of a square foot, plugging that into the average square-foot rate for real estate in the area where the customer lives would probably only amount to perhaps a only a few additional pennies per month.

You are referencing wrong rates; you are referencing residential real-estate rates, but Comcast is using the real-estate for a commercial purpose, and when you rent out a small bit of real-estate for a commercial purpose, the expected rates are higher than personal usage. Lookup colocation rates for 1U of rackspace in low-tier data centers. Comcast is colocating a modem, which is comparable to colocating a 1U router. Obviously, you don't expect them to pay for delivering a service to you, but if they are using their colocation to generate revenue by taking advantage of the prime location of your property to deliver revenue-generating services outside your customer relationship, then you are entitled to a share of the extra revenue that placement on your property is used to generate independent of their usage to deliver your service.

For example, to declare even a *portion* of your rent or mortgage as a business expense in a home business you have to actually almost *exclusively* dedicate some square footage of your home, such as a den or what have you, to that business, and not use it for any personal purposes

This is only true if you are both the owner of the home and the owner of the business. And it is nothing more than a rule designed to prevent self-dealing on your taxes where you claim some rent to be a business expense without actually sacrificing anything to the business. The IRS rules also have some differences from the actual law, and you could challenge them.

Since the 2nd modem is virtualized, it should not affect your transfer rates or bandwidth quotas.

WiFi can be fairly CPU intensive, and it can interfere with your own emitters.

This kind of interference is much more a problem than a neighbor's WiFi, as WiFi is low powered unlicensed emission, therefore a little bit of distance and walls tend to reduce the noise; this can increase the noise, resulting in worse throughput for wireless LAN clients in your home.

Unless there's a vulnerability in the router (which is possible)

Possible? It is almost certain. We just don't have the details yet.

Do you have a solution that doesn't require caller ID service?

ATT wants an arm and a leg to have caller ID svc.

why would there still be an additional public IP on the modem?

Just because your service is being bridged, doesn't necessarily mean that the modem isn't acting as a router for other services.

It can also be assigned an additional public IP outside the forwarding plane for your service for management purposes.

Plenty of reasons for a device made to act as a bridge to still have an IP. Also, seeing as its DOCSIS, the additional IP it has if any can be completely discrete and not discoverable by the subscriber of the bridged service

I would think,that comcast subscribers who have their modems used in this way should probably receive a monthly rebate on their bill of 50 cents

Nice try.... the customer should also have fair compensation for some other things of value that are being used by Comcast to generate these extra revenue for Comcast.

Mainly, the use of the customer's real-estate which the customer pays for and pays taxes on, for a purpose not related to service delivery to the customer, and in a manner which generates noise into their local RF environment on frequencies the customer may wish to use.

Next; bandwidth from the customer's service drop, which may affect their speeds if public users have high usage. Finally, the concern about additional potential risks for customer's network and service stability.

I can think of a few ways of disabling it.... most of them involve the use of a screwdriver and a pair of snips, or placing the modem/router in a shielded metal box.

I realize Vyatta can reach 10-Gigabit, but this is in a lab with idealized 1500byte or large packets, not real-world traffic which comes in a lot of shapes and sizes, especially during an attack, and then we have protocols which are highly latency sensitive or re-ordering sensitive such as VoIP, and "tricks" to try and reach 10-Gigabit throughput are compromising jitter-sensitive protocols.

We use Linux on our routers and it works just fine (we have about 400Mbit traffic on our AS).

How do you know for a fact that it works just fine? Are you quite certain that it really is 400Mbit, and not 800Mbit of traffic that inconsistently performing routers are quashing at peak time? :)

Have you hooked up a Spirent avalanche to a port on the router in a lab and sent 10-Gigabits of IP datagrams formulated as 70 Byte packets with randomized payload to randomized destinations transmitted at 17.8 Million packets per second, and measured a packet loss of 0% on the next hop?

Or, are you relying on the fact that the TCP stacks on your customers' equipment detect the increase in latency on your Vyatta router during congestion and automatically scale back their usage, causing download times to increase, while praying nobody uses Bittorrent, or gets infected and sends a UDP flood, or uses other more aggressive protocols and notices artifacts caused by being linked up through an ISP not using fully non-blocking (wirespeed) equipment?

You will have an IS degree + years of experience + an MBA. There s a large amount of career potential in that.

Or go for a Doctor of Business Administration. Anyways... I thought the snazzy thing these days was "MBA Equivalent"; you know... like getting a Masters in Finance and taking some other random classes. So you are slightly differentiated from the standard 'MBA' curriculua :)

But telling someone with no Cisco training to "Go and get a CCIE" is like telling a year 12 student to "Go and get a PhD".

Well..... I think there is one good reason to tell this to someone who has no certifications and has stated that they think certs aren't worth it: It should be a humbling experience, and hopefully they won't get to the point of blowing stupid amounts of money on an exam they can't possibly expect to pass. :)

What the vendor calls the device is more about what the primary intended place for the device is on the network; it is a reflection on the "standard" configuration of the device, or at least the defaults.

The problem with this assessment "Switches use ASICs; routers don't" is it is only true for low-end devices. The only way you are getting away without ASICs is if you are in a small-business, branch office, or Enterprise WAN with little traffic.

Switches require ASICs, but "routers" need them too. Routers typically need more advanced ASICs, since they need to look at Layer 3 network prefixes, not just a simple list of MAC addresses for L2 bridging.

The Cisco ASR routers use ASICs extensively, ditto for 76xx routers; in fact, they are exclusively used for forwarding, there is very little or no software-based switching through a high-end router. If a condition occurs where you run out of hardware TCAM or lose CEF and revert to non-ASIC-based software switching, it will be a very bad day indeed.

Juniper M/T/MX series edge routers are the same way. All forwarding is done in a separate ASIC-based hardware forwarding plane. Packets are not interpreted or forwarded by software. Even firewall rules, QoS policies, etc, are handled by ASICs on a reasonably high end router.

Once upon a time there were cases where you needed to upgrade PBB cards or policy feature cards on routers to add to policy management/access list functionality. These are definitely hardware-driven functions.

Common Layer 3 switches in fixed access configs have similar capability in some respects but more limited featuresets and limited capacity for table sizes, typically; you often don't have quite the same IP policy management features as on a full blown router; some of the L3 switches don't even have decent QoS (which is terrible).

Also if you need to take a full BGP table; you are not going to want to use a fixed-configuration Layer 3 access switch to do that --- since it probably lacks the memory, and even if it had the memory, generally there will be no supervisor redundancy.

The requirement to support a huge IP forwarding table, which requires extra memory and CPU, is what an Edge router needs on a multihomed network.

So there are clearly devices that specialize in being better edge routers than switches.

Almost all switches are L-3 switches, and almost all routers will bridge ports, so is there a difference?

A Layer 3 switch is just another word that means the exact same thing as router.

In fact.... the world's very first router ran on what is now an ancient microcomputer, and it was called a packet switch

Usually, when a vendor has multiple product lines and they describe some products as Layer 3 Switches and some products as Routers; what they are actually telling you is what primary task the device has been designed for. Something called a "Router" is a device designed primarily for a specific layer 3 duty, usually WAN, Edge, or Core routing IP services.

For example: Typically, when Cisco calls something a switch or Layer 3 switch, what this means is that the device is primarily designed for Layer 2 switching with light or "simple" Layer 3 duties, with limited flexibility (As in, the ASICs, TCAM, forwarding table sizes, and policy feature support will be smaller or partitioned differently by default than in a router).

It is probably great for Inter-Vlan routing at the edge and forming adjacencies with other Layer 3 switches and the core routers over common routing protocols such as OSPF, and with an advanced license, it may sometimes provide more "advanced" features such as BGP, however, the CPU and memory capacity available in the Layer 3 switch will not be designed for the heavy core router duties.

This is only a rule of thumb.... in some cases they are so close, that it is practically meaningless to distinguish between the "Router" and the "Switch".... the Cisco 76xx routing platform versus 68xx switching platform comes to mind.

However; the differences in the design of the specific devices specifications, and featuresets available tend to be more pronounced in fixed-configuration devices.

In the modular devices, you can pretty much use the modules to configure router as switch and vice-versa, so "L3 Switch" and "Router" have no real difference in that case

These are valuable certifications. They are also (I suspect) entirely out of reach for the OP for now, probably, seeing as he's in his 30s and has not pursued even the lower level certs. The thing is, these certs are somewhat progressive, and you need to have some experience preparing and taking certification tests before you take the slightly harder tests like the CCIE written, which is still, I understand, a cakewalk compared to the IE lab.

You don't just wake up one day and decide to sit the CCIE; you build-up to that level.... most candidates would earn a CCNA, then a CCNP; the NA is really rough for folks who have little experience in IT or little experience with certs, or who don't know binary like the back of their hand, the NA then needs 2 to 3 years of experience before they are ready to consider tackling NP, then require an additional 18 to 24 months after you earned that CCNP, working extensively with the technology building up experience, having little time for a family or social life, and dedicating practically every other waking minute outside work to preparing for the extremely rigorous lab.

I believe the formal description is "certification track". The RHCE, the CCIE, the JNCIE, or the VCDX is the pinnacle. Those are the certs you want to put on a resume. The others exist as milestones.

Once you achieve those milestones, you know you have been heading in the right direction, which is why you don't ignore those milestones.

If you pass the CCIE troubleshooting and configuration lab, the in-person interview, and the written test, then you can truly call that an accomplishment, and it will truly set your resume apart.

Listing any of the other certs is just saying "I'm one out of the other X million peopl on that path of development who can't yet put down a pinnacle cert."

These days, we have high quality software already made, And the open source revolution has shown us that it is poppycock, the claim that you need the proceeds of extortionate licensing fees in order to build high quality software.