How long can intel maintain such a large process lead over rivals?

[Lonbjerg]

why hate?

That ping just means he is close to the test server.

Most of those speed test sites are useless. You are only testing your speed to the test site is doesn't equal the speed you will see over the whole internet.

Oh no...accordingly to some people on this site, lantency is not a fucntion of distance....but of switching/routing! :whiste:

Quite hillarious...and I agree with your...sites like speedtest.net are not good tools.

 

[ShintaiDK]

Oh no...accordingly to some people on this site, lantency is not a fucntion of distance....but of switching/routing! :whiste:

Quite hillarious...and I agree with your...sites like speedtest.net are not good tools.

LOL. If latency wasnt a function of distance we would have solved all latency issues in thw world now

 

[Ferzerp]

Oh no...accordingly to some people on this site, lantency is not a fucntion of distance....but of switching/routing! :whiste:

Quite hillarious...and I agree with your...sites like speedtest.net are not good tools.

Well, it is both. Distance (travelled, not by the bird's flight) sets the floor, and routing and switching just add to that (either very small amounts (almost trivial amounts), or very large in the case of congested, broken, or misconfigured equipment).

Then there is the fiber vs copper argument that people don't grasp. Signal propagation is nearly the same (somewhere around .7c if I remember properly) in both mediums. Fiber does require substantially less power though (we're talking at 10Gbit speeds).


Steam gives me typically 6.5MB/s downloads. The speedtest number is not anomalous.

 

[Lonbjerg]

LOL. If latency wasnt a function of distance we would have solved all latency issues in thw world now

I know...even on a geek site people can surprise me with being uninformed

Well, it is both. Distance (travelled, not by the bird's flight) sets the floor, and routing and switching just add to that (either very small amounts (almost trivial amounts), or very large in the case of congested, broken, or misconfigured equipment).

Then there is the fiber vs copper argument that people don't grasp. Signal propagation is nearly the same (somewhere around .7c if I remember properly) in both mediums. Fiber does require substantially less power though (we're talking at 10Gbit speeds).


Steam gives me typically 6.5MB/s downloads. The speedtest number is not anomalous.

Yup, light moves a ~0.65 x c in a fiber (light only moves a c in a vacuum) and the signal propagation in a copperwire is ~0.65 x c too.

But if i look at my route to anandtech...the lantency is ~1ms due to switching/routing...the rest of the 137 ms lantency is pure distance. going from DK -> NL -> UK -> US.

 

[Ferzerp]

I know...even on a geek site people can surprise me with being uninformed



Yup, light moves a ~0.65 x c in a fiber (light only moves a c in a vacuum) and the signal propagation in a copperwire is ~0.65 x c too.

But if i look at my route to anandtech...the lantency is ~1ms due to switching/routing...the rest of the 137 ms lantency is pure distance. going from DK -> NL -> UK -> US.

Well, it's tiny in comparison, but to say that it doesn't contribute at all is an oversimplification. It's a smaller fraction for long distances, but it is a greater fraction for smaller distances. Envision a college campus with 3 or 4 hops within a mile or less of distance. In the case, while the latency is very, very low, the fraction of that latency attributable to the equipment and not the transport medium is a larger fraction than say, a 15 hop transit halfway around the globe.

 

[Lonbjerg]

Well, it's tiny in comparison, but to say that it doesn't contribute at all is an oversimplification. It's a smaller fraction for long distances, but it is a greater fraction for smaller distances. Envision a college campus with 3 or 4 hops within a mile or less of distance. In the case, while the latency is very, very low, the fraction of that latency attributable to the equipment and not the transport medium is a larger fraction than say, a 15 hop transit halfway around the globe.

There is a difference..the claim was that switching/routing was the major reason for lantecy...on normal customer based internetlines, not backbones

There switching takes a very small 3rd place...after protocol lantency and distance

 

[Ferzerp]

Major, no. contributor, yes.

You could argue that the contribution is so small that it can almost be ignored, but saying it has *no* impact is incorrect, just as saying it is the largest contributor is incorrect.

The speed of light is too danged slow. .7 of the speed of light is *WAY* too slow. I'd love to do synchronous data replication across about 1200 miles but can I? No, because light is too danged slow.

 

[dagamer34]

I know...even on a geek site people can surprise me with being uninformed



Yup, light moves a ~0.65 x c in a fiber (light only moves a c in a vacuum) and the signal propagation in a copperwire is ~0.65 x c too.

But if i look at my route to anandtech...the lantency is ~1ms due to switching/routing...the rest of the 137 ms lantency is pure distance. going from DK -> NL -> UK -> US.

Ideally, a website would have a CDN that places a server hosting AT content closer to you, but that costs quite a bit of money.

 

[Lonbjerg]

Major, no. contributor, yes.

You could argue that the contribution is so small that it can almost be ignored, but saying it has *no* impact is incorrect, just as saying it is the largest contributor is incorrect.

The speed of light is too danged slow. .7 of the speed of light is *WAY* too slow. I'd love to do synchronous data replication across about 1200 miles but can I? No, because light is too danged slow.

I never said it had no impact..everything does.
But as you say...it (in a working setup) should be so low that is virtually a none-issue.

 

[cbn]

Not without competition it won't happen.

The Intel of yesteryear used to make only Microprocessors. That's it. They've since delved into SSDs, Mobos, NICs, Graphics and even designed products like the Ultrabook concept as well as that TV rumor. The next logical step for Intel, assuming they can get away with the backlash, is to wrap all of this up in plastic and put an Intel logo on both sides of the product and not just a small sticker. I guess it's just a matter of money. If the OEMs pushing their own products would potentially mean less $$ for Intel than making the products and competing with the OEMs themselves then I'm sure they'd do it. Unlike Apple, Intel controls pretty much everything from the design > fab > product minus the plastic whereas Apple does it from the plastic backwards.

This is why I think Intel will try to innovate rather than play the "Me Too" game with Apple.

If Intel innovates sufficiently I think they would not need the business of wrapping commodity items (LCD, battery, Plastic or metal case) around their electronics and software.