Forum

I'm wonding if any of you have implemented WiFi coverage for 16,000 seat indoor basketball stadium and wouldn't mind sharing your thoughts. We all know about what they look like inside - a court, and in our case, four stories of seating for the fans.

Everything from the second to the fourth seating area is solid steel with chairs and benches welded in place, which for the most part eliminates installing access points in the concourse under the seating area because the signal will not penetrate into the seating area.

There is a large catwalk that criss-crosses the stadium above, with speakers and large screens for the fans to watch. My first thought was to use some sort of directional antenna on the catwalk aiming down at WiFi zones. I'm concerned about co-channel interference, etc. The signal will reflect all over the place during the survey and implementation, but will be different when 16,000 fans are in there to absorb the RF.

We already have 1000+ access points on our campus and two controller farms and think that utilizing them is the best, most cost effective way to go about this since we have plenty of growth potential built in.

Thanks in advance,
Tim

The question is, I guess, what kinds of clients do you have and what services do you need to provide? Do you have point of sales terminals that need to be serviced all over the stadium? What kind of devices are they? Do you have Wifi users? Where will they be located? What is their density? Do you have the media show up and need spectrum for the blimp, wireless cameras, high-bandwidth uploads for pictures etc? Is there a pressroom where they will periodically need guest access? Is there any existing wireless? You might want to do a site survey and find out if anything you don't know about is in place. After that you need to find out where you need coverage. Next, you need to find out where you can mount access points and where can you get power/data to those APs. In addition, any chance you have to use mesh to provide coverage/backhaul will save you time, cabling and effort. Don't assume you will have to provide wireless coverage to 16,000 users, because 90% of the time, that is not true.

We're a large university, so you never know what types of clients we'll end up with. We have existing WiFi in there, we deliver a student SSID, Guest SSID, ticket scanner SSID and stats monitor SSID. So far, no iPhones. No POS terminals - they're wired for PCI compliance. So to answer your question, mostly laptops, ticket scanners, and anything that might come into play in the future.

The clients will be in the seating area. Hard to believe people will bring their laptops to a game, though.

Students camp out before a game on a first come, first serve basis. They'll camp out with their laptops for a day or two in advance just to get tickets.

As for density, if you build it, they will come. Right now, mostly press and scanners, so less than 200 devices on the existing Wifi that works well for its intended purpose. But that is not in the court, just near the doors and press seating area.

No blimps, since this stadium has a roof. No wireless cameras - the press uses wired for that.

The press room already has guest access and is covered.

I've surveyed it several times in the concourse area around (behind) the seating area. The existing WiFi works well in that area. I'm only interested in the court and seating area at this time.

I was thinking we would have to run some fiber optic cabling up to the catwalk and install a PoE switch and run a dozen access points to far reaching areas and use directional antennas. I'm concerned about reflections since the entire place is steel.

And who really know how many people will even have WiFi devices. My experience so far is we design for one purpose and they always end up wanting to use it for another...

Most definitely you should deploy 11n AP's since they love multipath.

It may sound counter-intuitive, but you probably don't want to deploy the AP's in the open arena. The problem is channel planning. Let's take 2.4 GHz for example, if you deploy more than 3 AP's in a space where they can hear each other on the same channel, throughput will suffer.

I'm speaking without seeing the building, but if it is possible to deploy the AP's behind the stands in some way it would be best. It is an outside in approach vs. inside out.

What brand of AP's are you using now? If it is Aruba or Cisco that limits you a bit.

If you were open to using a different vendor, it could open up some possibilities.

Aerohive - Excellent product with great meshing capability without a controller, but RF management is key for your deployment.

Xirrus - Nice array, but I don't think I'd recommend it for such a large area. I've never deployed it though so I'm going from what I've read.

Meru - Advantage of good transmission coordination and could work well.

Ruckus - I'm on a Ruckus kick right now since I'm doing the roadshow, but I have seen the technology do some amazing things. True beam forming (vs. chip based) has major advantages and I know it would work well for your situation. Mesh would be a huge plus.

Beware of the over saturation of AP's. Some vendors are notorious for using too many AP's, especially in situations like this.

I did a bit of Googlevestigating and it looks like you are at UK. I'm not too far away and would like to see what you have going on up there. Anyway, sounds like a fun project!

GT

See inline, GT...

Most definitely you should deploy 11n AP's since they love multipath.

My reply: We deployed 802.11n in the concourse around "the outside" of the basketball court and I have found that roaming doesn't work that well. Possibly because the chipsets in the clients aren't on the same page with the controller. Maybe that will iron out now that 802.11n is ratified? I was actually going to turn high throughput off because the a/b/g clients are working beautifully there.

It may sound counter-intuitive, but you probably don't want to deploy the AP's in the open arena. The problem is channel planning. Let's take 2.4 GHz for example, if you deploy more than 3 AP's in a space where they can hear each other on the same channel, throughput will suffer.

My reply: Agreed. But I could use 5GHz... And I was thinking of using some directional antennas shooting down from the catwalk, but am curious if anyone else has done the same. I think that during the game there will be less multipath since there will be 16000 people in there.

I'm speaking without seeing the building, but if it is possible to deploy the AP's behind the stands in some way it would be best. It is an outside in approach vs. inside out.

My reply: Agreed. I tried that since we have APs on the first floor concourse. Between floors of the concourses is a 10 inch slab of concrete. Between the 2nd and 3rd floor concourses and the seating area is on big slab of steel. And on the inside, the seats are metal as well and they're welded on to the steet. I guess steel was inexpensive in 1955.

What brand of AP's are you using now? If it is Aruba or Cisco that limits you a bit.

My reply: We already have a mix of 1000+ Cisco 1230s, 1240s, 1250s and 1140s. For the most part they do a fine job for us. Sure would be nice to not have to buy another backend to support another vendor's equipment. As you can guess, our backend wasn't cheap.

If you were open to using a different vendor, it could open up some possibilities.

My reply: Agreed.

Aerohive - Excellent product with great meshing capability without a controller, but RF management is key for your deployment.

My reply: I was thinking of using Cisco's mesh, actually. Have never deployed it, but have built a dozen P2P links using Cisco autonomous products in the past witch great success.

Xirrus - Nice array, but I don't think I'd recommend it for such a large area. I've never deployed it though so I'm going from what I've read.

My reply: Me too -- I've read about it, held it in my hand, but have never deployed it or even powered it on.

Meru - Advantage of good transmission coordination and could work well.

My reply: Would that mean buying another backend?

Ruckus - I'm on a Ruckus kick right now since I'm doing the roadshow, but I have seen the technology do some amazing things. True beam forming (vs. chip based) has major advantages and I know it would work well for your situation. Mesh would be a huge plus.

My reply: Don't suppose that road show will stop in Kansas City... And does that mean another backend? I can imagine clients would want to roam from inside the arena to the concourse where the Cisco APs are.

Beware of the over saturation of AP's. Some vendors are notorious for using too many AP's, especially in situations like this.

My reply: I've heard the same stories. I had several ideas in my head. Utilize the existing backend(we're not made of money), use mesh, maybe use 1250s with patch antennas with high throughput turned off, and with any of the above thoughts, turn the 2.4 GHz power down.

I did a bit of Googlevestigating and it looks like you are at UK. I'm not too far away and would like to see what you have going on up there. Anyway, sounds like a fun project!

My reply: If you're even in the Kansas City area, let me know!

Tim,

Of course the Cisco gear will work fine if properly deployed in that environment.

The directional antennas off of the catwalk could definitely work. I would be curious to see what kind of coverage overlap vs channel planning could happen. Does Cisco have band steering? If not, we would create two SSID's, one for 2.4 GHz and another for 5 GHz. 2.4 GHz would be "Jayhawk" and 5 GHz would be "SuperSonicJayhawk" or something like that.

Is roaming something that people will use? Of the original post, it seems like much of it was static usage.

When are you looking to deploy?

GT

GT,

Good guess on the "jayhawk"... you hit that one right on the head.

I'm not sure when deployment will happen. It depends on price, actually. I'll get a number together and they'll either go with it or not.

As for roaming, you never know. The ticket scanners roam sometimes from door to door. I have noticed that roaming doesn't happen like I would like when clients are using 802.11n.

In fact, the 2.4GHz would most likely be 802.11b/g and 5GHz would be for 802.11n.

Thanks for the reply.

Tim

Not sure what you mean by 2.4 GHz for 11b/g and 5 GHz for 11n. Unless there is a reason I'm not aware of, you want to use 11n for both, even if you are supporting legacy STAs. With 11n, you get the advantages of MRC and better signal transmission from the APs.

GT

I have sat through countless webinars and listened to "the experts" say that you should keep 2.4GHz for 802.11b/g and use 5GHz for 802.11n. One reason is if you use 40 MHz channels you basically use up the 2.4GHz spectrum unless you only have one access point in the building.

I have installed 100+ 802.11n access points in the last year, and I must admit I'm less then impressed with roaming on 802.11n. I have seen clients roam on those same access points when all they have is 802.11b/g adapters, but the 802.11n clients just don't work properly.

For that reason alone I am thinking of not enabling high throughput in the 2.4 GHz spectrum. After all, it is my phone that rings when there are problems...

Now I understand. When I say 11n AP's, I don't necessarily mean in HT mode. 11n AP's in non-HT (b/g) mode have superior communication with legacy devices.

My initial comment was to caution people not to deploy 11 b/g AP's just because they will only be supporting b/g STA's. 11n AP's, in any mode with any STA's have superior communication skills.

So your 11n STA's aren't roaming properly? Have you had time to investigate that problem? STA problems are the worst because there isn't usually much that can be done...

GT