The 49ers' plan to build the greatest stadium Wi-Fi network of all time
- Thread starter JournalBot
- Start date
I am rather curious to see what kind of new apps this will lead to. Also, yay if I can order a gourmet hot dog and Coke to my seat!
Upvote
11
(12
/
-1)
Great article! I don't know much about networking but small, cohesive, and motivated teams get me excited about any subject.
I'll have to buy some tickets when the new stadium is up.
I'll have to buy some tickets when the new stadium is up.
Upvote
14
(15
/
-1)
I run a small 5 ap network and had trouble keeping the network from going to a crawl during the hurricane. I cant imagine how they will do it for a whole stadium. We use the dlink corporate ones . I also have 3 brocade ones sitting here we got for cheap that are normally $1k each. That wifi network for the 49ners will be expensive.
Upvote
0
(2
/
-2)
A little off-topic, but I've noticed that plays (in college football at least, I couldn't care less about pro) are just over 30 seconds apart typically. That makes for very nice skipping when watching a game on DVR. Watch play, skip 30 seconds, next play is set up. You can watch the entire game in less than 30 minutes usually.Article":3mwnvbdl said:
Upvote
8
(8
/
0)
So why go to the extra effort to limit the wireless connection speed to the wide receiver sprinting towards the end zone? (sorry couldn't resist!)Article":1skdqfyp said:
Upvote
-2
(3
/
-5)
Would be great to see some of these techniques come to WWDC or other conferences.
I've never been at a busy conference where the wifi worked reliably at a keynote or other packed event.
I've never been at a busy conference where the wifi worked reliably at a keynote or other packed event.
Upvote
10
(11
/
-1)
Azethoth666":3gsw72oz said:
If it's anything like MLB, you can order it, but you have to go pick it up. Only the more premium seats bring it to you.
Upvote
7
(7
/
0)
Upvote
26
(27
/
-1)
This is an awfully long article to basically say, "We're building a super-awesome wireless network that will solve problems nobody has been able to solve, but all we can tell you is that we're doing all the same things any competent network designer would do."
Reads more like a fluff piece from an airline magazine.
Reads more like a fluff piece from an airline magazine.
Upvote
24
(33
/
-9)
I think the problem will not be coverage during the game, but during any delays or breaks in action. While fans are in their seats, many won't be using wifi (based on the Patriots experience). However, during halftime or during a long delay (weather, injury, replay) fans will be out of their seats, in walkways, and using wifi. I feel like the walkways should have significantly more coverage that seating areas.
Upvote
3
(3
/
0)
I'm interested to see how well 5GHz works for stadiums. If the FCC ends up approving more spectrum in the 5GHz range, the 49ers could have a lot more than just 8 or 9 20MHz channels..
(from a NTIA report)
U-NII-1 is indoor only, U-NII-2 and 3 are indoor/outdoor, however U-NII-2 is subject to dynamic frequency selection for airport radars.
(from a NTIA report)
U-NII-1 is indoor only, U-NII-2 and 3 are indoor/outdoor, however U-NII-2 is subject to dynamic frequency selection for airport radars.
Upvote
5
(5
/
0)
Zero technical detail. Physics is physics, and building a wireless network that deals with 2.4 Ghz co-channel interference is nothing like Facebook building a data center where you can just add more bandwidth.
Pure fluff.
Pure fluff.
Upvote
37
(39
/
-2)
neuromaster":2zuzmctm said:Three pages later I'm still wondering what their plan is
While there isn't a lot of meat to the article, there are a few tidbits hidden which assist in their general plan.
Excluding how the wired parts of the network work, placing access points under the stairs between sections on the way down (instead of just at the top and bottom), set to single channels (so there is little possible contention between neigbouring access points) and then tuning the transmit power on each ap (turning it down) to decrease overlap, would enable one to squeeze more access points into a smaller area and thus increase the number of total possible connections.
Yes there are issues to be resolved to deal with throttling, where to best put antennas, etc.. but it isn't impossible, just difficult.
Upvote
8
(8
/
0)
You forgot the part where you only have 3 non-overlapping channels to work with, assuming 2.4 Ghz is still the most widely used band (like anything before an iPhone 5, Nexus 7, etc.). Unless they plan on installing a parabolic reflector under each seat, there still be co-channel interference. RF doesn't just stop propagating on a strictly defined boundary unless you build a bunch of Faraday cages, so there will be interference.
The whole point is that talk is cheap. This article is all sizzle and no steak. Unless they invented or are doing something new (which the article doesn't mention) it is as believable as your friendly neighborhood wireless vendor account manager. I am sure it will be as good as the money they throw at it will make it - no more and no less.
The whole point is that talk is cheap. This article is all sizzle and no steak. Unless they invented or are doing something new (which the article doesn't mention) it is as believable as your friendly neighborhood wireless vendor account manager. I am sure it will be as good as the money they throw at it will make it - no more and no less.
lyme":jxz10oi4 said:
Upvote
9
(11
/
-2)
Anyone else notice the iOS dominance in Gillette Stadium's WiFi network (see included screenshot of their network dashboard app), mirroring all the other webstats we get from Ars on that type of thing? Android users don't go online with their phones? It's not hard to connect to a new WiFi network on Android, is it?
Upvote
6
(9
/
-3)
As a point of clarification, is this free wifi? Or is there some sort of a paywall required for accessing it?
Are there intranet pages that lead you to buy stuff? I assume there's some sort of a goal here, and they're not just upgrading the network out of the goodness of their hearts.
Are there intranet pages that lead you to buy stuff? I assume there's some sort of a goal here, and they're not just upgrading the network out of the goodness of their hearts.
Upvote
2
(3
/
-1)
Stone":2e4wgv6a said:
Nothing is free. The seat licenses for the new stadium are not priced for the faint of heart.
Upvote
13
(13
/
0)
It's difficult to understand the context of some statements in the article. Simultaneous association/connectivity and concurrent throughput are two very different conversations. That "conventional wisdom" is based upon real world limitations. Wireless radio frequencies, transmit and receive airtime have actual limitations. These limitations are even more apparent with 1x1 mobile devices. Some very bold statements made and promises made.
Upvote
10
(10
/
0)
colo":9uzdm31p said:
While what you say is likely true, it doesn't jive with what the article discusses.
Upvote
2
(6
/
-4)
http://prod.static.49ers.clubs.nfl.com/ ... ingMap.pdfkray28":15v8xag6 said:
I know stadium tickets are expensive, and there's probably a fat service charge on top of the ticket prices listed. I'd just like to know if this is a premium purchase over the price of a seat.
It may be that they're touting a system that will never come close to the design usage. How many people carry wifi-only devices anymore? I know that cell data is probably shitty in a location this crowded, but the actual services their wifi provides would be interesting info to have in conjunction with this article.
Upvote
1
(2
/
-1)
While I'm not too much of a football fan, I drive by this "stadium in progress" at least once a week, and this article, along with watching the stadium coming together is generally piquing my interests.
Upvote
0
(1
/
-1)
Krutawn":1miu852e said:
If what they said was true, that shouldn't be a problem anyway. They claim that they will be able to support simultaneous use from every person in attendance at LTE speeds. While that's hard to believe, it's still unlikely that every person will be connected at once.
Also, if you do a little math on the numbers he gave: terabit of throughput / 68,500 people comes out to just under 2MB/s per user.
Upvote
3
(3
/
0)
lyme":2ipimpbd said:
2.4Ghz is still the predominant technology. There are only 3 non-overlapping channels as colo points out. Unless the article has omitted that the stadium intends to drive towards the 5Ghz band only, the numbers don't add up. Even with 5Ghz only, I would argue that co-channel interference remains a real discussion.
Some stadiums mentioned seem to be headed the right direction. I would want to see the empirical data that supports those 49ers claims. It's interesting that the Superbowl saw less than a season game at Gillette. I noticed the only screen shot is from NetSight (Enterasys product). Was there a less technical savy crowd at the Superbowl?
Upvote
2
(2
/
0)
I'm calling bull on the no outsourcing. You think those ten guys are running all that cable?
Upvote
6
(8
/
-2)
Turning off channel bonding in the 5Ghz space in high density areas is not a panacea, nor is it uncommon. It makes sense when you don't need quite as much bandwidth, and you are more concerned about co-channel interference. Remember, channel bonding is a relatively recent feature for 802.11 (2009 - 802.11n). 5Ghz is great if all your devices support it, but that is a big "if" and not solvable by the infrastructure.
I am still waiting on some specifics on what they are doing that is unique. I sure it will be swell, but the technical detail is lacking.
I am still waiting on some specifics on what they are doing that is unique. I sure it will be swell, but the technical detail is lacking.
lyme":2cq0vc74 said:
Upvote
1
(1
/
0)
Upvote
-3
(0
/
-3)
As @colo says, the article is a lot talk. In fact, to me it sounds barely above a marketing hype session.
Bandwidth isn't a problem for a deployment like this - enough money tossed at it and it's not an issue. The issue is density. WiFi has a severely limited density factor to it. You can deploy a couple thousand APs in close proximity, but that doesn't mean they'll have clear access to the backhaul supporting it - to get to that, you still have to deal with a clear channel to the AP, with enough spectrum time to transmit and receive. With so few channels in the 2.4 range and so much AP density, transmission slots are going to be *HIGHLY* curtailed. They may not be doing bandwidth caps, but they'll get it anyway by nature of the spectrum usage!
The solution is a multipronged approach - utilizing apps that will minimize chatter and do a lot of localized caching, and hardcore spectrum analysis from all sorts of places around the stadium. They seem to be banking on 5 ghz channels, but that's only going to get you so far.
As for the comparisons to Facebook and a datacenter, those are all ridiculous. The only similarities they have in common are that FB and datacenters use lots of bandwidth and to isolate networks they'll be using a domain-ified network. Who cares? That's barely an issue. The problem at the stadium is WiFi connectivity. No datacenter uses WiFi for anything but low bandwidth, non-mission critical connections or for the handful of people who are in the area. If you're talking about the FB office (the FB datacenter falls under the same umbrella as any other datacenter), I seriously doubt their office infrastructure is as dense as a football stadium, and that like most of us who work at an office, we'll plug in our ethernet ports when the WiFi gets annoying.
Bandwidth isn't a problem for a deployment like this - enough money tossed at it and it's not an issue. The issue is density. WiFi has a severely limited density factor to it. You can deploy a couple thousand APs in close proximity, but that doesn't mean they'll have clear access to the backhaul supporting it - to get to that, you still have to deal with a clear channel to the AP, with enough spectrum time to transmit and receive. With so few channels in the 2.4 range and so much AP density, transmission slots are going to be *HIGHLY* curtailed. They may not be doing bandwidth caps, but they'll get it anyway by nature of the spectrum usage!
The solution is a multipronged approach - utilizing apps that will minimize chatter and do a lot of localized caching, and hardcore spectrum analysis from all sorts of places around the stadium. They seem to be banking on 5 ghz channels, but that's only going to get you so far.
As for the comparisons to Facebook and a datacenter, those are all ridiculous. The only similarities they have in common are that FB and datacenters use lots of bandwidth and to isolate networks they'll be using a domain-ified network. Who cares? That's barely an issue. The problem at the stadium is WiFi connectivity. No datacenter uses WiFi for anything but low bandwidth, non-mission critical connections or for the handful of people who are in the area. If you're talking about the FB office (the FB datacenter falls under the same umbrella as any other datacenter), I seriously doubt their office infrastructure is as dense as a football stadium, and that like most of us who work at an office, we'll plug in our ethernet ports when the WiFi gets annoying.
Upvote
11
(11
/
0)
I'd love to get a follow up article with more hard details. I understand that they are still in the planning phase and not even really in implementation but the only real detail they provided is that they seem to be banking on 5ghz devices being much more common so they can move a lot of devices there to avoid overloading 2.4ghz for devices that only support that. 5ghz wifi is inherently better for this type of deployment because it has more available channels, with even more coming eventually from the looks of the recent FCC changes, and it doesn't propagate as well through walls etc so it's easier to avoid interference.
Upvote
2
(3
/
-1)
Erm . . . about to actually read the article, but I think the 49ers in your second paragraph should be NFL champions, not NFC champions.
Upvote
-16
(1
/
-17)
Its great to talk the talk, but without including any details whatsoever, that's really all it is.
It sounds more like "At this time, we don't think 68,500 people will ever actually try and connect and then pull 20-40Mbps and hopefully by the time that's really necessary we'll beyond 2.4GHz restrictions."
Also 68,500 people pulling 20-40Mbps are going to saturate their devices before your 1Tbps pipe, or did I just misread that?
It sounds more like "At this time, we don't think 68,500 people will ever actually try and connect and then pull 20-40Mbps and hopefully by the time that's really necessary we'll beyond 2.4GHz restrictions."
Also 68,500 people pulling 20-40Mbps are going to saturate their devices before your 1Tbps pipe, or did I just misread that?
Upvote
1
(1
/
0)
Why are so many people in these comments ignoring the key point in the article?
These guys essentially admit that the only reason they will be able to accomplish what they're talking about is because of the industry move to 5GHz. And that they wouldn't be able to do it if they had to deal with today's mix of cell phones (i.e. not enough people currently have 5GHz-capable phones).
That's essentially the whole story: Why will the 49ers be able to provide 100% Wi-Fi coverage when no one else has? Because it'll be a year-and-a-half from now, and technology improves over time.
The article sort of implies no one else is working on 100% coverage in their stadiums for that kind of time frame. I guess that's possible. Maybe the 49ers just have an advantage because they happen to be rolling out a whole new network at exactly the same time that 5GHz will allow for 100% coverage.
Setting up the stadium network at level 3 instead of level 2 is also sort of interesting. But face it, backhaul isn't really the constraint for serving a stadium full of people with Wi-Fi, so that's just an interesting optimization.
These guys essentially admit that the only reason they will be able to accomplish what they're talking about is because of the industry move to 5GHz. And that they wouldn't be able to do it if they had to deal with today's mix of cell phones (i.e. not enough people currently have 5GHz-capable phones).
That's essentially the whole story: Why will the 49ers be able to provide 100% Wi-Fi coverage when no one else has? Because it'll be a year-and-a-half from now, and technology improves over time.
The article sort of implies no one else is working on 100% coverage in their stadiums for that kind of time frame. I guess that's possible. Maybe the 49ers just have an advantage because they happen to be rolling out a whole new network at exactly the same time that 5GHz will allow for 100% coverage.
Setting up the stadium network at level 3 instead of level 2 is also sort of interesting. But face it, backhaul isn't really the constraint for serving a stadium full of people with Wi-Fi, so that's just an interesting optimization.
Upvote
11
(11
/
0)
I'd say it's highly improbably that a stadium full of people would all be connecting to the wifi network at all, let alone all at the same instant. So these delusions of grandeur are pretty much risk-free at this point.
Upvote
1
(3
/
-2)
I think the bandwidth numbers gloss over that there's an assumption that a big chunk of bandwidth demand will be video served off of servers at the stadium. That bandwidth never hits the external pipe.wkhtl":2xxg6a7f said:
Upvote
8
(8
/
0)
Upvote
4
(4
/
0)
- Status