Welcome!

Microservices Expo Authors: Pat Romanski, Liz McMillan, Elizabeth White, Stefana Muller, Karthick Viswanathan

Related Topics: Microservices Expo, Java IoT, Industrial IoT, Microsoft Cloud, Machine Learning , Agile Computing

Microservices Expo: Article

100 Years in the Movies: One Evening’s Web Performance

Why one company performed better during this year’s Super Bowl

Both Paramount and Universal celebrated their 100th anniversary last year, which is a long time to be in the movie business. Arguably, both have made some good, some great, and some bad movies. But, during this year's Super Bowl, Paramount showed Universal how to design a ‘fast and furious' web site that stood up to the flood of visitors during and after the game.

This article will discuss not only how Paramount was able to do it, but will also compare Universal and Paramount's Super Bowl web site results, which shines a light on key factors for successful web performance: fewer connections to fewer hosts requesting smaller objects produces a smaller page size having a positive impact on page response time.

To begin, Universal and Paramount are near equals when it comes to their web age. Jumping over to http://web.archive.org, I found Paramount launched its first site 16 years ago in 1997 and Universal's first site came online 15 years ago in 1998. With the same amount of experience on the Web, it's interesting to explore why one company performed better during this year's Super Bowl.

As a result of analyzing web page performance for nearly six years, my experience tells me that the reasons some sites succeed and others don't can fall into three general categories - corporate culture, resources, and experience and knowledge. But even today, with so much information available on web site performance fundamentals, I often see companies forgetting the basics.

Why Universal Was not ‘Fast and Furious'
Looking at Paramount and Universal's site performance for the period from 5 p.m. EST until 11 p.m. EST on Sunday, February 3 (Super Bowl Sunday), I noticed some big performance differences between the two sites. For starters, Paramount's homepage average response time was 966 milliseconds while Universal's was 11.727 seconds.

Comparison of the two sites clearly shows that the differences come down to web performance basics and the fundamental construction of a web page. This includes connections, object count and type, page size, and hosts. This can be seen in the following chart:

†3 objects greater than 200KB

*11 objects greater than 200KB

The Fewer Connections the Better
Paramount designed their Super Bowl site using only nine connections, while Universal used 41 connections:

The number of connections was a significant factor in Universal's poor response time because more connections equated to more bytes transferred. The tradeoffs can be significant as Universal's response time shows.

After the IP address is resolved by a DNS lookup, the number of connections generally sets the pace for page loading. Even though modern browsers are capable of making between 6 and 8 simultaneous connections to the same host, it doesn't mean you have to use them all. Creating TCP/IP requests takes time and resources, and the milliseconds of overhead that each one requires can quickly add up to seconds, especially during a big game night when massive web traffic is expected.

Identify Flying Objects
The amount of time spent making HTTP requests for all the objects can have a marked impact on page response time. The following chart shows that Paramount's homepage contained only 41 objects, while Universal's homepage contained 121 objects:

Once a connection is established the objects will, presumably, just fly into your visitors' browser, right? Not always.

What if these objects are large files and the servers are straining because of an increased load (as during a special event like the Super Bowl). In Universal's case, there were 11 files tipping the scale at well over 200KB each (two files where over 750KB each). Paramount, on the other hand, only had three files exceeding 200KB.

Size Matters
You've probably heard this once or twice - page size matters. Page size is characterized by totaling all the files that make up a web page (typically compressed and in KB).

Here we see Universal's homepage size coming in at a whopping 4995KB while Paramount's homepage comes in at only 1625KB:

Typically, file size isn't too much of an issue during normal surfing, but Super Bowl Sunday is not a normal traffic day. You can agree that five pounds is heavier than two pounds, and it subsequently takes more effort to lift five pounds. This same concept is true for web sites - some are heavy in KB and others comparatively lighter.

In this case, Universal's page was not as ‘fast and furious' as Paramount's because its page size was 3370KB heavier. Newtonian Law of the Internet states it's going to take longer to download heavy pages than lighter ones so long as the access lines are equal.

Host Counts Count
Using the HTTP Archive Trends site (http://httparchive.org/trends.php#numDomains&maxDomainReqs), you can find information on many web site design trends. Two such trends that I find interesting are the average number of domains accessed across all websites and the maximum number of requests (Max Reqs) on the most used domain.

Here's a table comparing Paramount to Universal to the average for all websites:

The difference between the two studio sites is very clear. We can see that Paramount designed a site that was well under the average for number of Domains and Max Reqs on one domain with three and 39, respectively, where Universal was well above at 27 and 75, respectively:

Further, examining the average number of Domain/Hosts alone, Universal used 9x more hosts across their homepage over Paramount. A fast response time can be increasingly challenging to design and there will be some compromises made, and a low host count seems to be an obvious tactic to follow.

Back to the Basics
Comparing the Universal and Paramount Super Bowl web site results highlights some of the key truths of web performance. Primarily, fewer connections to fewer hosts requesting fewer, smaller objects produces a smaller page size and these items can have a positive impact on page response time. Placing these areas on high alert before going live - connections, object count and type, page size, and hosts - may be one of the best ways to ensure a successful Super Bowl any day of the year.

More Stories By Gregory Speckhart

Gregory Speckhart is a Senior APM Solutions Consultant at Compuware.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


Microservices Articles
The explosion of new web/cloud/IoT-based applications and the data they generate are transforming our world right before our eyes. In this rush to adopt these new technologies, organizations are often ignoring fundamental questions concerning who owns the data and failing to ask for permission to conduct invasive surveillance of their customers. Organizations that are not transparent about how their systems gather data telemetry without offering shared data ownership risk product rejection, regu...
Containers and Kubernetes allow for code portability across on-premise VMs, bare metal, or multiple cloud provider environments. Yet, despite this portability promise, developers may include configuration and application definitions that constrain or even eliminate application portability. In this session we'll describe best practices for "configuration as code" in a Kubernetes environment. We will demonstrate how a properly constructed containerized app can be deployed to both Amazon and Azure ...
DevOps is often described as a combination of technology and culture. Without both, DevOps isn't complete. However, applying the culture to outdated technology is a recipe for disaster; as response times grow and connections between teams are delayed by technology, the culture will die. A Nutanix Enterprise Cloud has many benefits that provide the needed base for a true DevOps paradigm. In their Day 3 Keynote at 20th Cloud Expo, Chris Brown, a Solutions Marketing Manager at Nutanix, and Mark Lav...
The now mainstream platform changes stemming from the first Internet boom brought many changes but didn’t really change the basic relationship between servers and the applications running on them. In fact, that was sort of the point. In his session at 18th Cloud Expo, Gordon Haff, senior cloud strategy marketing and evangelism manager at Red Hat, will discuss how today’s workloads require a new model and a new platform for development and execution. The platform must handle a wide range of rec...
The Internet of Things is clearly many things: data collection and analytics, wearables, Smart Grids and Smart Cities, the Industrial Internet, and more. Cool platforms like Arduino, Raspberry Pi, Intel's Galileo and Edison, and a diverse world of sensors are making the IoT a great toy box for developers in all these areas. In this Power Panel at @ThingsExpo, moderated by Conference Chair Roger Strukhoff, panelists discussed what things are the most important, which will have the most profound e...
If your cloud deployment is on AWS with predictable workloads, Reserved Instances (RIs) can provide your business substantial savings compared to pay-as-you-go, on-demand services alone. Continuous monitoring of cloud usage and active management of Elastic Compute Cloud (EC2), Relational Database Service (RDS) and ElastiCache through RIs will optimize performance. Learn how you can purchase and apply the right Reserved Instances for optimum utilization and increased ROI.
TCP (Transmission Control Protocol) is a common and reliable transmission protocol on the Internet. TCP was introduced in the 70s by Stanford University for US Defense to establish connectivity between distributed systems to maintain a backup of defense information. At the time, TCP was introduced to communicate amongst a selected set of devices for a smaller dataset over shorter distances. As the Internet evolved, however, the number of applications and users, and the types of data accessed and...
Consumer-driven contracts are an essential part of a mature microservice testing portfolio enabling independent service deployments. In this presentation we'll provide an overview of the tools, patterns and pain points we've seen when implementing contract testing in large development organizations.
In his session at 19th Cloud Expo, Claude Remillard, Principal Program Manager in Developer Division at Microsoft, contrasted how his team used config as code and immutable patterns for continuous delivery of microservices and apps to the cloud. He showed how the immutable patterns helps developers do away with most of the complexity of config as code-enabling scenarios such as rollback, zero downtime upgrades with far greater simplicity. He also demoed building immutable pipelines in the cloud ...
You have great SaaS business app ideas. You want to turn your idea quickly into a functional and engaging proof of concept. You need to be able to modify it to meet customers' needs, and you need to deliver a complete and secure SaaS application. How could you achieve all the above and yet avoid unforeseen IT requirements that add unnecessary cost and complexity? You also want your app to be responsive in any device at any time. In his session at 19th Cloud Expo, Mark Allen, General Manager of...