Google's Dark Fibre Plans?
sebFlyte writes "According to news.com "Google is looking for Strategic Negotiator candidates with experience in...(i)dentification, selection, and negotiation of dark fiber contracts both in metropolitan areas and over long distances as part of development of a global backbone network." Is the search giant planning to build a global fibre-optic network?" Or perhaps simply use unused fibre that they can get for cheaper then from the datacenter providers; although at least from my talks with the datacenter folks, Google's not paying much per Mbps as it is.
Ah, so has Google unified String Theory, Dark Matter, and Dark Energy?
"Absorbing your worst..."
BTW, the Light Reading guys were the ones who "broke" this story back on January 6th
Hulk SMASH Celiac Disease
And thus it begins....it was difficult to see at first; what Google's plans were. Only after it had struck first using highspeed fiber to initiate the subroutines in the Google desktop search companion did humanity realize its vast mistake. Only it was too late...Google was selfaware...and it was hungry...for pie...I mean Pi...
Did you know you can be apathetic to apathy? Not that I give a shit...
Give him a break, he's only an editor. It's not like he's supposed to know the difference. Oh, wait...
The AMS-IX is the largest Internet Exchange / NAP in Europe.
This is your sig. There are thousands more, but this one is yours.
Maybe this is the beginning of Google Broadband. With all the other non-search areas they've gotten themselves into, maybe they're looking to take on MSN and Yahoo in the ISP realm.
GoogleISP: Dark fiber to your city, fiber to you home coming soon.
And after they can give everyone a super high speed broadband connection, it's just one more step to selling a subscription for the comping suite of web-based apps that GMail proves they're so good at.
They've got several (lots of?) datacenters that have to sync up lots of data. Anyone with enough data to transfer around can save money just buying the strands of fiber or wavelengths on lit fiber instead of paying a provider to light it. It's not surprising that Google has enough of this work to do that they want to hire someone with experience in it.
Either that or they want to be able to sync their world-wide collection of data-servers in an 'accpetable' length of time.
A hollow heart and empty head makes the streets run red.
I would be. It's far from their core competency and there's so much competition in the telco business that everyone sells at cost anyway. Maybe a datacenter chain like Equinix would be a better acquisition target.
Which word is "dark" modifying, "plans" or "fibre"?
--
so dark, you'll forget the fibre
No no no.
Why the hell would Google want to buy up an existing ISP/telco with all the crap that that entails?
What they are doing is actually very sensible.
By looking to negotiate purchase/lease of dark fibre over the medium term they are avoiding the big cost which is actually putting fibre into the ground.
I imagine that they would ensure that the maintenance of that fibre is the responsibility of the provider, so they don't need to run their own maintenance crews either.
And the BIG plus with having access to fiber is that you can then ramp up your capacity by using WDM (Wave Division Multiplexing) to get more bandwidth out of your fibre.
They have probably realised that to ramp up their networks to cope with their future plans they need more bandwidth that they can afford to buy as "service" from a regular telco. Its just too damn expensive!
By leasing the fibre themselves, they light it how they want, rather that how the telco wants to sell it to them.
This *might* have biogger up front costs, but the recurrent costs are MUCH lower.
Do you have any statistics to prove what you said that "The AMS-IX is the largest Internet Exchange / NAP in Europe."?
Last I heard the largest Internet Exchange was located in London.
cat
And even where there is overcapicity, it is mostly in the urban areas, put in place for business, not single family homes. Good luck getting dark fiber in the 'burbs, let alone the sticks
"Well, good luck finding a judge that doesn't run a bestiality site."
It is "cheaper than", not "cheaper then". Only on slashdot does a Belgian reader teach English grammar to an American editor.
Am I the only one who thought of a few laxitive jokes when they read this?
"The price good men pay for indifference to public affairs is to be ruled by evil men." ~Plato (427-347 BC)
Both AMS-IX and LINX staff can impress the girls/guys with having the bigger one. LINX's switches are pushing more data (71 Gbit total include private+public/49.23Gbit for public peering) Ams-ix is doing 49.7 (dunno if that includes private)
;-)
AMS-IX has more members 210 against LINX having only 169 members.
However they are both equally fun to party with.
Use Adsense for Charity
We are practically a Google Temple here, folks.
The unofficial
Don't forget that Google parks hardware all over the place. They've got a pile of it sitting in datacenters run by Savvis, who bought up the dregs of the Cable & Wireless operations (who bought up the dregs of the Exodus operations). But unlike Exodus, C&W also had tons of dark fibre. Savvis has been trying to make everything lean and mean, but they've got a pretty nice inter-datacenter-network... but not necessarily any bargains when peering with other people's operations. I can imagine that Google would love to get outside the loop of having the datacenter operators dictate what terms they're willing to live with when setting up new peering arrangements. Especially as Google's needs become more instantaneously multi-directional (rather than crunch-and-publish, it's real-time ad stats, mail, etc).
Even if all these new hires do is help Google's datacenter providers make good decisions about new or altered peering networks, they'll probably earn their keep.
Don't disappoint your bird dog. Go to the range.
Hello,
Considering what Google has built internally for server management and redundancy, I would hypothesize based upon available data (i.e. GFS) that they're looking to light up fiber between their data centers, while running either TCP/IP or IPv6 (with modifications of existing IGP and routing protocols, more than likely BGP or OSPF) between them.
This is a very smart move on their part, if this is true. This would allow them to do their own internal traffic control and shaping over a private network, and develop/modify algorithms for efficient transfer of data over said network, without having to "play" by Telco/ISP rules.
In other words, they're more than likely building their own global network to more efficiently transfer data over the Internet by completely bypassing it for their inter-server traffic. This is a very smart move, if true.
FWIW it is very common for larger companies to buy up Dark Fiber from large telco providers. All of the Tier 1 Telcos sell Dark Fiber as a standard product line, and you will find many of the Fortune 100 companies out there own their own. I highly doubt this is Google's atempt to change business plans and enter the depressed Telco sector.
They're going to create a global super computer. Local servers gives higher performance. What do you think the Google bar is for? Coming versions will ask you if you want to donate your free CPU cycles to Google. Google in turn will sell these to corporations that needs CPU power.
One possible explanation is that Google is looking for something to do with their current $55B stock valuation. Other than making the founders incredibly wealthy, the high stock price by itself isn't particularly useful to the company. But it can be used as collateral for loans to acquire assets that could be useful both now and in the future. Given the massive storage and computing resources that Google already manages, I suspect that they can manage their own fiber network for very little incremental expense (I'm less sure about the physical care of the fiber -- who fixes your dark fiber when someone cuts it?). As they attempt to provide more and more services, they may simply want more control over the underlying transport.