Wireless Wednesday–"Super Bowl XLVIII: How did Tier 1 wireless carriers’ networks hold up?"
February 5, 2014 — by WEBOOST
This week we’re talking about the big game still for our “Wireless Wednesday” post. Take a look at the excerpt from the @FierceWireless article posted below–
Now that the game is over, wireless carriers are boasting about the traffic they saw at the venue and how their networks held up at the game, which more than 85,000 people attended.
In a company blog post, Verizon said it recorded record volumes of data two hours before kickoff. The carrier notched 800 percent more data connections during this weekend’s game than during the busiest hour at last year’s Super Bowl in New Orleans. Further, Verizon said that during the halftime show with Bruno Mars and the Red Hot Chili Peppers, Verizon used more data in MetLife stadium during a single hour than at any previous Super Bowl. Verizon said that the total customer in-stadium data usage on its network was 1.9 terabytes, which it said was a Super Bowl record.
As for AT&T, John Donovan said AT&T’s total data usage in the stadium on its network was more than 624 GB, which is the highest data usage AT&T had ever seen from a one-day sporting event it has measured. Further, AT&T said traffic peaked at 5-6 p.m. Eastern Time, which was the hour directly before kick-off, and during that hour 119 GB crossed its in-stadium network. That peak data usage is up 52 percent from the 78 GB during the busiest hour at last year’s Super Bowl.
Sprint said in a statement that its network “performed exceptionally well during the Super Bowl on Sunday, besting our competition in voice performance and providing consistent, solid data performance. Sprint saw 4G LTE data traffic increases over a typical game day of 83% and 150% for download and upload speeds respectively. And, Sprint was using a single channel of LTE, so about half of what our competitors are using today. With the deployment of Sprint Spark, Sprint will be able to aggregate multiple channels of LTE to boost throughput speeds that will match and ultimately exceed competitor speeds.”
T-Mobile said that compared to the New Orleans Super Bowl in 2013, data traffic increased by approximately 700 percent during the busiest hour for its network, but during that period it still saw peak download speeds reach over 60 Mbps. Additionally, T-Mobile said that compared to what it typically sees at MetLife Stadium, voice traffic during the Super Bowl increased by approximately 70 percent.
However, FierceWireless worked with Nexgen Wireless, an independent, third-party telecommunications software and engineering services vendor, to see how the carriers’ networks performed at the event.
Nexgen’s competitive benchmark studies focus on in-building traffic at major events. During the Super Bowl, Nexgen tested the carriers’ 3G voice and LTE data quality. The data was collected for each network using the vendor’s Accuver XCAL-M software loaded on the Samsung S III, with a handset for each carrier’s network. This information then was uploaded to a server for Nexgen’s team in Chicago to process and compile using Nexgen’s proprietary tools.
One very important caveat: Nexgen said its XCAL-M software was not ready to support T-Mobile’s handsets and file structure. Thus, T-Mobile’s network performance was not tested.
Each testing sequence for each carrier was conducted at six locations throughout the stadium. Six tests were done. The first was conducted 45 minutes prior to kickoff, the second was at kickoff, the third was at approximately 7:30 p.m. ET, the fourth was at halftime in the game, the fifth was 45 minutes after the start of the third quarter, or 9 p.m. ET, and the sixth was at approximately 9:45 p.m. ET.
Nexgen measured average downlink throughput, average uplink throughput, ping latency, voice call success rates ((total attempts minus failures)/total attempts), and voice call drop rates ((total successful calls minus drops)/total successful calls).
For downloads, the tests used a 20 MB file, and for uploads the tests used a 10 MB file. To measure latency, Nexgen conducted a total of 100 ping tests with packet size of 32 bytes for each sequence. A short voice call test consisted of 10 short calls (approximately 20 seconds in length each) and long call test consisted of three long calls (approximately 2 minutes in length each).
What do you think of this data? Let us know in the comments section below, or on Facebook/Twitter.
For buildings greater than 7500 sq ft, we recommend our WilsonPro brand of products.