IBM's Big Data story in the U.S. Open

I live in New York and I grew up here, so for me, this time of year has a very special meaning. Although I am not a very keen sports fan, but I like to watch (tennis) Open, the game in Queens, New York, Flushing Meadow Park. At the end of every summer since high school I was in the open or watching on TV. Back in the 1980s, my dad's company had a side box at the Armstrong Stadium (at the time of the central court) and every year our family went there to buy tickets and watch a few games in the fourth and One-fourth finals. So it's very exciting to be able to sit in the third row of the game and watch my favorite McEnroe, Jimmy Connas, Lendl, Billot Bog and others compete for the supremacy of tennis.

I've been following many of the characters who dominate the game. So for a while I focused on IBM. The technology IBM has built up in the open over the years has become increasingly complex. And, as this technology has become more and more accessible to individual enthusiasts, I want to know more about how it is built. So when I visited IBM's sponsorship marketing technology manager Kent This week, I was kind of like a kid in a candy store, and he told me the truth about the technical features and infrastructure of IBM's U.S. Open. Here I will give a full message of what I have gathered from the newsletter.

The U.S. Open on the ipad

IBM has been working with the US Tennis League, the behind-the-scenes organization of the US Open, which has lasted for 22 years now. Throughout the period, the company has been providing infrastructure to transmit scores and statistics, but in the past few years things have become interesting.

This year, for example, the US Open has its own ipad app (iphone apps implemented in 2009), breaking scoring information, live and on-demand video and analysis. Even from a complex social media point of view: When an ipad application is opened, you will see a set of columns, each showing a series of messages, which are hashed labels, specific to the open National Tennis Center of the individual Stadium (and competition).

With just a few taps, you can navigate the video stream's menu, either the pitch or the live game, of course, HD. You can also easily see the latest picture of all 5 major tournaments (men, women singles and doubles, and mixed doubles). This week, in the island rented beach house, I used the main app, where the digital subscriber line (below 3Mbps downstream) was the only broadband option, and the reader worked well.

Back to the network

However, you don't necessarily need an ipad to take advantage of this technology. For example, if you log on to www.usopen.org on your desktop or laptop, you can take advantage of many of the same feature programs, even some that you won't get on the ipad. For example, select a video and broadcast from the top navigation bar, then click the US Open Live from the Drop-down menu, select a course, and start watching. As I write this article, I'm watching Angeli Cober and Williams are playing their second leg. Since I'm on the internet, I don't just want to watch linear video and listen to all kinds of reviews; In the live video, I also want to do some superposition of data analysis. As you can imagine, IBM has not let me down.

First of all, I can click on the game Data button to see the real-time updated statistics, such as the percentage of the first serve, the number of service double errors and natural errors, and I can see these calculated numbers about the entire match or specific settings. In the upper-left corner of the screen, there is a button labeled "Enter the game button." "If I click on it, some very interesting data will appear." For the first time in last year's open program, three of the most important factors can be identified for a given match, tailored specifically for each particular competitor. On the site's SlamTracker feature, you can see the same data (not as an overlay on a live video) that can be used for fractions and statistics:

These measures also include the number of targets and states. In fact, they are key performance indicators (KPIs), back to competitive sports, the game shows the key is a scorecard, so that the term can be used in a full cycle. But not directly on-line Analytical Processing (OLAP), these key performance metrics are based on the execution forecast analysis of data from all four Grand Slam tournaments in the past 7 years, with a total of 39 million positions.

The final winner by analyzing the most popular cobre in most competitions. Although the Cobeby Venus has a higher ranking, the result of the race is indeed somewhat unsettling, as is IBM's accurate predictive analysis.

This is truly valuable data, and IBM is working with the Entertainment Sports television Network (ESPN) to keep the conversation going by providing the data to the latter's live-broadcast critic. But it is clear that in this network age and interactive analysis, you no longer need to rely on commentators to get it. Instead, you just need a browser and a mouse, or an ipad and your finger, to be your own grand Slam data analyst.

What is

hiding behind the scenes?

About two weeks ago, I wrote an article "IBM's Big Data capabilities", including its product portfolio and acquisitions. So naturally, I would like to know what the back end of the IBM product is and how its technology is used in the U.S. Open. Here's what I know:

SlamTracker Technology (including competition data and game key) is heavily using the SPSS technology acquired by IBM in 2009.


IBM has a powerful relational database, and the database is used very frequently for scoring data and operations.


WebSphere MQ (FKA MQ series), IBM based messaging based middleware is used for scoring delivery, allowing you to get online scores more quickly.


WebSphere Technology is used for the overall service architecture.

What interests me most is how much of the technology used in the above list can last more than 10 years (some of them are also more). Core statistics, relationships, SOA, and middleware technologies have not become unimportant at this stage of data and analysis. In addition, it is particularly noteworthy that the Hadoop, Netezza and Cognos Business intelligence technologies have not been cut. Hadoop, data warehousing, and business Intelligence (BI) are of course important, but the traditional enterprise technology of IBM applications shows that large data and BI-specific technologies are not necessarily prerequisites for good analysis implementations.

(Responsible editor: The good of the Legacy)