Google broadband

http://www.nytimes.com/2010/02/11/technology/companies/11google.html

Google is looking for towns with populations of between 50 and 500 thousand people to tweak their network infrastructure in order to deliver 1gbps Internet access to homes and businesses. They'll be spending hundreds of millions on this project and people who pay for the service will be paying for the normal cost of broadband.

Just to be clear, it's fiber to the home. So it's better than cable in two ways. It's a fat pipe--1gbps which is a ton of bandwidth, and it will be low latency, which is different from bandwidth and important for real time communication.

My guess is that this is going to be disruptive as hell. 1gbps to the home is unthinkable. What's going to come of this? My guess is that when Google announces which towns it will install the broadband into, floods of people will move to those towns for new jobs made available by the new networks. What would you do with 1gbps to your doorstep? Nevermind, I know what you would do. Besides high definition 3d porn, what would you do with all the bandwidth?

When I think of a huge pipe with a low ping, I think virtual reality. The ability to move massive amounts of user data back and forth in real time makes it possible. Couple that with the kind of shit we're seeing with motion capture it seems like a logical step.

 

Log in or Sign up to hide all adverts.

I have fiber to two of my homes already.

Only way to go.

Cost a lot of money to have it done in 1998. You could buy a Ferrari with that kind of money. At that time I could move more bandwidth out of my home than any home in America, according to the best of our investigative efforts involving major fiber providers.

Now it's being offered in quite a few cities at regular rates.

1gig is a screaming rate no doubt.

 

Log in or Sign up to hide all adverts.

Almost all servers limit request speeds to well below that level. I think a 15mbps connection will max out most connection speeds. It seems to be a pointless endeavor unless you're going to find a connection speed for servers thats even larger or somehow the entire internet goes p2p.

 

Log in or Sign up to hide all adverts.

Yep, I agree. No one is going to be able to get the full use of the system because the servers would not be able to handle it. Once this thing is in place, it will be the servers themselves that become the bottleneck - and a BIG one at that.

 

Yes I can tell you quite factually that I was faster than the internet in 1998 with 15 meg fiber link(fraction DS3). Way faster.

Today I have faster connectivity and I am faster than the internet.

But sending files within a city would be smoking fast. There would be times for such a thing. And multitasking or business operations would be cool.

Someone could open a local online file backup on a link like that and really have a great service for local customers with that kind of link speed.

I'm glad someone is moving us forward in the US though. It's common to have 100 meg+ connections in other countries, they are ahead of us.

 

Yes, I'm also glad to see it. At least it's a *good* step in the right direction.

Aside from server farms having to grow considerably in throughput, the primary problem that will remain for some time is getting bigger pipes (fiber) to the millions of homes and businesses that are located outside the greater metro areas. And the FCC is due to announce/implement a plan to do just that very shortly. Although, instead of fiber at first, it will come in the form of wireless.

 

I've been thinking about this concept off and on now. Even though internet speeds might not significantly improve, intranet speeds would improve. That is, the city of 50k would essentially become a large intranet. Google's MO is about integrating internet application into common society...it makes sense for their goal. If they can build a small infrastructure, they would have a huge testing pool for intranet applications they might develop with a more community-based feel.

 

One of the articles I was reading was saying that the backbone carriers are up to date with speeds and it's the local providers who are lagging behind. So if you have a number of cities who have 1gbps ftth Internet access, then it's more than a city-wide superfast intranet. You'd be able to send files across the country at 1gbps from one house to another. It's the local carriers who are the bottleneck, which is what Google is trying to change.

 

Bring some real service and competitive pricing where it’s NEEDED — we’ve dealt with regional monopolies for far too long in this country…. and if they want to start charging tiered pricing, they can bite it — with Google coming in and offering GBPS for a fair price & supporting Net Neutrality, the other demons are toast.

Imagine the gamers faces after reading the Google news for the first time, Salivating! Even little things like calling your loved ones on the phone and seeing each other in real time on a screen/monitor with HD+ clarity would be a big deal to many families.The list is endless.

Please Register or Log in to view the hidden image!

 

That would totally depend on what technology/topology they are planning to use when they deploy (IE - something infamous for being strained by customer-congestion like a "token ring").
They may very well have a good system in the works, and the ability to deliver. The articles I've seen so far only mention that Google wants to build a test network, but no mention of specifics.

Also - 1 gig subscriber line is a lot by today's standards, but that stuff is all relative. It's kinda like the fact that 256 meg flash memory was something special a few years ago......and nowadays, you can practically get 4 gig drives at the bottom of a box of cracker jacks.

 

It's sort of relative and sort of not. 1gbps is a giant leap. It's going to allow us to do many new things, things we may not have imagined. And it's not going to be in some isolated city like others have suggested. Google is going to tweak the networks in a small number of cities across the US. So the US is going to have a number of dots on it that have extreme data transferral ability between each other, kind of making them like hubs for certain types of applications like telemedicine.

 

You could sit inside a ball of 3d monitors each displaying "Google Earth real-time stereoscopic edition" whilst downloading Heroes, Stargate, Prisonbreak, Naruto, Big Bang... all at the same time and have them before lunch.

 

I don't believe Google can pull this off alone. I think it's just a tactic used to spur other internet providers to up their game. And it makes perfect sense.

Faster internet = more pages surfed = more google ads shown = $$Google$$

The hype from Google's announcement alone probably made them millions.

 

expanding on read's server comment........

at what speeds can data be currently written into the different kinds of storage media out in the market....and at what cost?

 

For off the shelf hard drives, you are looking at:

Solid state: ~135 MB/sec
Traditional (10k RPM):

"As of 2008, a typical 7200rpm desktop hard drive has a sustained "disk-to-buffer" data transfer rate of about 70 megabytes per second.[41] This rate depends on the track location, so it will be highest for data on the outer tracks (where there are more data sectors) and lower toward the inner tracks (where there are fewer data sectors); and is generally somewhat higher for 10,000rpm drives. A current widely-used standard for the "buffer-to-computer" interface is 3.0 Gbit/s SATA, which can send about 300 megabyte/s from the buffer to the computer, and thus is still comfortably ahead of today's disk-to-buffer transfer rates. Data transfer rate (read/write) can be measured by writing a large file to disk using special file generator tools, then reading back the file. Transfer rate can be influenced by file system fragmentation and the layout of the files." - wikipedia


Processor speed limitations and multithreading will need to be considered as well.

 

Super High Def delivery.

 

High definition everything is a logical and incremental improvement. What I'm thinking is more along the lines of entirely new applications.

Haha... this is what I could see happening. Some dumb 20 year old punk decides, Hey, I'm going to do whatever I can to completely use up ]b]all[/b] of my bandwidth. And then somehow he manages to do it, and in doing so, he stumbles upon a few good ideas for real cool and useful applications.

But profiting from something like this is going to require partners--people who have a fat pipe on the other end--to make use of the great and obvious potential for new modes of real time remote communication.

 

It opens up even more possibilities for distributed computing as well.

 

The above is pretty good information.

But SCSI is still king of disc. When you want performance you go SCSI even over solid state. It's a shame that so few PC enthusiasts these days seem to be aware of SCSI, much less use it. It used to be de rigueur for performance seekers. Though certainly solid state is the future, after they work out the kinks with file sizes and TRIMing and the like. SCSI is the best there is still.

A 4 disc array of 15,000 RPM Seagate Cheetah 3 SCSI discs in a stripe configuration for performance, meaning all the discs read and write as one big, really fast drive, still smokes everything. The discs last longer, are tested to far higher standards than any serial(SATA) drives, and they can put out throughput of 130 megs a second off a single disc. Even a SATA RAID is much faster than a single disc.

Websites that offer up files: Google, Yahoo, Youtube, etc, use very fast RAID arrays with RAM(solid state) caches to just cook data out to the internet. Some of the solid state(RAM) caches are way up in the gigs now, so gigs of data goes out at RAM speeds before the RAID even comes into play.

What I'm saying is, a gig a second doesn't change all that much for a fairly modern PC as to disc in general. But SCSI is king, but somehow forgotten in many minds.

That Google connection will not connect to another city that fast, whoever posted that above is incorrect. Even our OC192(over 9 gigs a second) and faster long-haul fiber backbones get swamped these days. OC768(over 30 gigs a second) and the like is being implemented but is still not the norm. OC192 is the most common backbone speed. So 9-10 people on Google would swamp the connection leaving the city for one carrier. No chance fellas.

As I said above: intra-city stuff would be nice, but not extra-city. Though I'm sure it wouldn't hurt anything, the 'net would be crisp I bet.

 

FTTH (Fiber To The Home) isn't new idea. In the countries such as Japan and South Korea, FTTH is getting common. FTTH will be dominant in near future, so Google's story is not much attractive to me.