Click here to close now.

Welcome!

AJAX & REA Authors: Carmen Gonzalez, Liz McMillan, Elizabeth White, Pat Romanski, Adrian Bridgwater

Related Topics: Java, AJAX & REA

Java: Article

Performance as Key to Success

How Online News Portals could do better

What factors make you think a web page is good or not? What keeps on that page longer than others? On the one hand it is the content on the page and whether this content is of interest to you. On the other it is the velocity with which you can navigate through the individual pages. High-Speed internet and performance-optimized pages make our day-to-day browsing easier when accessing our emails, tweets, latest updates on sports or news. With all the changes in the recent years in Web Performance Optimization (WPO) we’ve been spoiled by those sites that follow all these Best Practices and boosted their web site experience. No wonder that we start losing our patience with a site that doesn’t respond as fast as we’ve come to expect.

For this particular reason Google modified their ranking algorithm to now also include web performance as one of their metrics. The speed of a web page – the time from entering the URL in the browser until the page has been fully loaded – has now become one factor that defines whether your site shows up on the top or on the bottom of a search result. A good reason to pick “fast by default” as the main theme for this year’s Velocity 2010 Conference. Steve Souders – the driving force behind Web Performance Optimization – hosted this conference and it became clear the significant impact Web Site Performance has on end user behavior and thus the success of a web site.

At last year’s Velocity the first results on Business Impact of Performance were presented to the public. Microsoft, Yahoo and Google presented their results of their internal tests that showed that slower pages have a direct impact on banner clicks – and therefore generated revenue. On the contrary – Shopzilla presented the results of a web site overhaul showing how a faster web site positively impacts the number of users, time spent on site, number of clicks and generated revenue.

Who is faster? Who is better? Bild.de or Spiegel.de?

Similar to my analysis of the FIFA World Cup Website, the Golf Masters or the web site for the Winter Olympic Games in Vancouver I wanted to take a closer look at the two biggest German news portals: Bild.de and Spiegel.de.

Based on Alexa, these are the top two news portals in Germany. Both portals probably have a very different reader community as both publishers have a different way to “present” news. My analysis focuses on how well these pages follow the Best Practices on Web Performance Optimization – therefore the actual content doesn’t matter, as it is all about how to deliver the content. Both pages deliver high-volume, multimedia content. Both pages make money with online ads and therefore it is in their interest that many users visit their site, stay as long as possible on their site and click as many ads as possible. We’ve learned from Google, Bing, etc… that speed is a critical factor in keeping user on the page – now let’s see how these two sites do:

There are three great and free tools available that we can use for an analysis like this:

All three tools analyze individual web sites based on the Best Practices from dynaTrace, Yahoo and Google and provide a nice overview on whether the rules discussed in these documents were met or not.

The Test scenario
On both sites – www.bild.de and www.spiegel.de – I looked at two individual pages: the start page and the overview page for politics. Important for my testing is to start with a cleared browser cache in order to analyze the page as a First Time Visitor. It is recommended to run the same test again – this time with a primed browser cache – and then compare the timings.

The Result
Using any of the three mentioned tools is rather simple. dynaTrace AJAX allows me to record a complete browser session which includes all 4 pages I am testing. The following screenshot shows the Performance Report that is opened when double clicking on the recorded session showing the four tested pages:

dynaTrace Performance Report shows all 4 pages in performance comparison

dynaTrace Performance Report shows all 4 pages in performance comparison

We can make the following interesting observations:

  1. Both sites have a First Impression Time of less than 2 seconds. This means that it takes less than 2 seconds until the user gets a first visual impression of the website – which is considered acceptable
  2. The Fully Loaded Time of both start pages is very high – both take about 14 seconds to fully load. The main reason for this is the amount of multimedia content (mainly images)
  3. Bild.de requires 289 HTTP Requests to fully load – that’s 113 more than Spiegel.de
  4. Bild.de uses JavaScript and XHR (XmlHttpRequests) to dynamically load content
  5. The full page size of bild.de is 4.6MB. That is 2.5 times more than Spiegel.de which “only” has 1.8MB

Lifetime of a Web Page
One of the nicest features of dynaTrace AJAX is the Timeline View. This view shows all activities (HTTP Requests, JavaScript execution, Rendering, XHR Requests and page events) that happen on a single page in chronological order. We can also see all requests split up by domain which makes it easy to spot which domains serve a lot of content or which domains serve the content very slowly. Especially on pages that include external 3rd party content, e.g.: Ad-Banners, it is interesting to see how this type of content slows down your page. Google’s PageSpeed has a new feature in their latest version which allows you to focus on your own or external content when analyzing page performance.

The following two screenshots show the dynaTrace Timeline for the start page of Bild.de and Spiegel.de:

Bild.de loads most of their images from a dedicated image domain. Rendering activity seems to very high as well

Bild.de loads most of their images from a dedicated image domain. Rendering activity seems to very high as well

Spiegel.de serves all images from its primary domain. It also hardly uses JavaScript or excessive rendering

Spiegel.de serves all images from its primary domain. It also hardly uses JavaScript or excessive rendering

There are some fundamental differences between the two start pages:

  • Bild.de uses multiple domains to deliver multimedia content, e.g.: bilder.bild.de or newscase.bild.de. Splitting content on multiple domains is a Best Practice that is called Domain Sharding. It has the advantage of letting the browser use more physical network connections to download more content in parallel
  • Spiegel.de delivers most of its multimedia content from the primary www.spiegel.de domain which is the bottleneck of their deployment. A browser has a limited number of physical network connections to a single domain, e.g: IE 7 uses 2 connections. If there are more than 2 resources to be downloaded from a domain they need to get queued up and have to wait for a connection to become available. Domain Sharding solves this problem by splitting content on multiple domains
  • both pages have individual resources that take extraordinarily long to load, e.g.: initial HTML page and CSS files on spiegel.de or 2 Flash components on bild.de
  • Bild.de shows a constantly high rendering activity – caused by animated images as well as the large number of images on that page

Key Performance Indicators
Additionally to the performance metrics as shown in the dynaTrace AJAX Performance Report Overview the Key Performance Indicator (KPI) tab shows a set of additional important KPI’s:

Key Performance Indicators for Bild.de - showing high load time, server-time, wait time and number of requests

Key Performance Indicators for Bild.de - showing high load time, server-time, wait time and number of requests

As mentioned earlier, bild.de requires many HTTP Roundtrips (289) in order to load the full page. The number one rule in Web Performance Optimization says to minimize HTTP Roundtrips. 289 roundtrips from browser to server is definitely too much. Every roundtrip needs to wait on a free physical network connection (read more on The Two HTTP Connection Limit Issue). Every roundtrip includes the overhead of network latency between browser and server as well as the overhead of the HTTP Protocol itself (HTTP Headers very often contribute a large percentage of the total roundtrip size).

The KPI Report also shows how many resources (images, CSS, …) are being cached or have to be retrieved from the server every time. The first table shows that 80 resources do not use any browser cache headers. That means that these resources have to be re-downloaded on every subsequent visit of the same user.

Usage of the Browser Cache
The Browser Caching Tab on the dynaTrace AJAX Performance Report performs a detailed analysis of HTTP Cache Headers on every downloaded resource. It is recommended to read Best Practices on Browser Caching which explains available browser caching options.

Bild.de and Spiegel.de don’t make optimal use of the browser cache. The following illustration shows a list of all resources that have no cache setting, a cache setting with a past date or a very short cache expiration date:

Many resoures on the page will not be cached at all or have a very short expiration date

Many resources on the page will not be cached at all or have a very short expiration date

Obviously it doesn’t make sense to cache every image on a page. Images that have a short life time should not fill up the local browser cache unnecessarily. If we have a closer look at some of the images on these pages here it seems that at least some of them could be cached for a longer time as they won’t change that frequently. The following section of the website shows images that are cached only shortly (< 48 hours):

Some examples on images that can be cached longer as they won't change that frequently

Some examples on images that can be cached longer as they won't change frequently

These logos will probably not change every 2 days – therefore it makes sense to specify a Far-Future Expiration Header. This reduces the number of roundtrips for revisiting users.

Reducing HTTP Roundtrips
I talked about this several times now – the top rule is to reduce network roundtrips. Besides making use of the browser’s cache there are other ways to reduce the roundtrips for every user (not just revisiting). dynaTrace AJAX has the Network tab that shows those roundtrips that are considered “unnecessary”:

7 unnecessary HTTP Redirects, CSS, JavaScript and Images on this page

Seven unnecessary HTTP Redirects, CSS, JavaScript and Images on this page

HTTP Redirects allow implementing some important use cases, e.g.: authentication, short memorable URLs or end-user monitoring. Too often though HTTP Redirects are caused by wrong configuration settings on the web server. Redirects that can be avoided are a great way to improve web site performance. On the start page of spiegel.de we have 5 redirects – bild.de has 7. A redirect means an additional HTTP Roundtrip as the browser needs to follow the redirect and request another URL in order to get to the originally requested resource.

Great savings can also be achieved on CSS, JavaScript and image resources. All 3 types can potentially be merged into fewer resources. The Best Practice on Network Requests and Roundtrips talks about CSS and JavaScript merging and compression. It also talked about CSS Sprites. CSS Sprites is a technique to combine multiple images in a single resource and use CSS styles to show the individual images on the correct location on the page. Check out the Best Practices on CSS Sprites.

Reducing resources not only brings the advantage of fewer roundtrips. It also means that fewer resources need to be downloaded from the same domain which reduces waiting time (remember the limitation of 2 connections per domain on IE7 mentioned earlier in this blog?).

On all pages of the two sites (Bild and Spiegel) we can observe too many JavaScript, CSS and image resources. It is not possible to merge all of these files – as this is technically not always feasible. It looks like some of these files could be merged and would therefore greatly enhance load time.

Dynamic content
Even though the majority of the content on a news site is static for at least a while (new news doesn’t arrive every second) there is enough content that needs to be generated dynamically for every user. Examples for this are any type of ticker information, weather or personalized ads.

dynaTrace identifies dynamic content based on the rules defined in Best Practice on Server-Side Performance. All these requests are listed on the Server-Side tab in the Performance Report. The Server-Time is the time also known as Time-To-First-Byte. It is the time from the last byte of the HTTP Request sent until the first byte of the response is received. This also includes network latency between the browser and the server – it is however the closest you can get to the server-time without actually measuring the time on the server. The following illustration shows the dynamic resources of bild.de:

dynaTrace shows slow running server-side requests - both on bild.de as well as Ad-Service domains

dynaTrace shows slow running server-side requests - both on bild.de as well as Ad-Service domains

The slowest requests are stock information requests, the initial page request itself as well as the weather information. We also see some requests to an external ad-service and requests to a web-tracking service (not visible in the screenshot above). Server-side performance mainly becomes an issue for dynamically generated content – and usually only in a scenario where many users want to access data from the server. In this case it becomes an even bigger problem because it prevents a larger number of users to actually click on those ads that generate revenue.

Top 10 Performance Problems taken from Zappos, Monster, Thomson and Co lists the typical server-side performance problems and shows how to prevent them.

Conclusion: Who delivers the fastest News?
In our tests, neither site does particularly well. Both sites deliver lots of content without following the Best Practices on Browser Caching or Network Roundtrips. The fully-loaded time of both pages is very high with 14 seconds each (obviously this also depends on your connection speed). Both sites deliver an acceptable first visual impression (<2 seconds).

dynaTrace, PageSpeed and YSlow allow uploading performance results to ShowSlow. ShowSlow is an Open Source Platform and can be used as performance repository for web metrics. ShowSlow.com hosts a public instance of the ShowSlow server and allows you to upload and compare results:

Performance comparision between the sites using all available Web Performance Tools

Performance comparison between the sites using all available Web Performance Tools

The difference in ranking and grading is that every tool puts the focus on different rules. dynaTrace puts its focus on load time first and then on rules such as browser caching or network roundtrips. YSlow and PageSpeed focus more on the Best Practice rules. A good mix of tool usage is therefore recommended – especially because no tool supports every browser anyway.

Now – who is the winner of this analysis? Let’s hope it is us – the readers of these online news portals. With the result of this analysis, with the help of the tools and with the help of people like Steve Souders, Web Site Performance is put in the spotlight and will ultimately lead to better and faster websites:

Performance == More Users == More Revenue

Related reading:

  1. How better Caching helps Frankfurt’s Airport Website to handle additional load caused by the Volcano Along with so many others I am stranded in Europe...
  2. Hands-On Guide: Verifying FIFA World Cup Web Site against Performance Best Practices Whether you call it Football, Futbol, Fussball, Futebol, Calcio or...
  3. Video on Common Performance Antipatterns online [caption id="attachment_1020" align="alignleft" width="250" caption="Parleys - Common Performance Antipatterns"][/caption] Last...
  4. Thomson Reuters on How they deliver High Performance Online Services We have had some great Webinars with our customers in...
  5. www.utah.travel in minutes">How to analyze and speed up content rich web sites likes www.utah.travel in minutes One of my daily activities is checking interesting blog posts...

More Stories By Andreas Grabner

Andreas Grabner has been helping companies improve their application performance for 15+ years. He is a regular contributor within Web Performance and DevOps communities and a prolific speaker at user groups and conferences around the world. Reach him at @grabnerandi

@CloudExpo Stories
17th Cloud Expo, taking place Nov 3-5, 2015, at the Santa Clara Convention Center in Santa Clara, CA, will feature technical sessions from a rock star conference faculty and the leading industry players in the world. Cloud computing is now being embraced by a majority of enterprises of all sizes. Yesterday's debate about public vs. private has transformed into the reality of hybrid cloud: a recent survey shows that 74% of enterprises have a hybrid cloud strategy. Meanwhile, 94% of enterprises a...
Thanks to Docker, it becomes very easy to leverage containers to build, ship, and run any Linux application on any kind of infrastructure. Docker is particularly helpful for microservice architectures because their successful implementation relies on a fast, efficient deployment mechanism – which is precisely one of the features of Docker. Microservice architectures are therefore becoming more popular, and are increasingly seen as an interesting option even for smaller projects, instead of bein...
Move from reactive to proactive cloud management in a heterogeneous cloud infrastructure. In his session at 16th Cloud Expo, Manoj Khabe, Innovative Solution-Focused Transformation Leader at Vicom Computer Services, Inc., will show how to replace a help desk-centric approach with an ITIL-based service model and service-centric CMDB that’s tightly integrated with an event and incident management platform. Learn how to expand the scope of operations management to service management. He will al...
In this scenarios approach Joe Thykattil, Technology Architect & Sales at TimeWarner / Navisite, presented examples that will allow business-savvy professionals to make informed decisions based on a sound business model. This model covered the technology options in detail as well as a financial analysis. The TCO (Total Cost of Ownership) and ROI (Return on Investment) demonstrated how to start, develop and formulate a business case that will allow both small and large scale projects to achieve...
The truth is, today’s databases are anything but agile – they are effectively static repositories that are cumbersome to work with, difficult to change, and cannot keep pace with application demands. Performance suffers as a result, and it takes far longer than it should to deliver new features and capabilities needed to make your organization competitive. As your application and business needs change, data repositories and structures get outmoded rapidly, resulting in increased work for applica...
The Workspace-as-a-Service (WaaS) market will grow to $6.4B by 2018. In his session at 16th Cloud Expo, Seth Bostock, CEO of IndependenceIT, will begin by walking the audience through the evolution of Workspace as-a-Service, where it is now vs. where it going. To look beyond the desktop we must understand exactly what WaaS is, who the users are, and where it is going in the future. IT departments, ISVs and service providers must look to workflow and automation capabilities to adapt to growing ...
The speed of product development has increased massively in the past 10 years. At the same time our formal secure development and SDL methodologies have fallen behind. This forces product developers to choose between rapid release times and security. In his session at DevOps Summit, Michael Murray, Director of Cyber Security Consulting and Assessment at GE Healthcare, examined the problems and presented some solutions for moving security into the DevOps lifecycle to ensure that we get fast AND ...
Since 2008 and for the first time in history, more than half of humans live in urban areas, urging cities to become “smart.” Today, cities can leverage the wide availability of smartphones combined with new technologies such as Beacons or NFC to connect their urban furniture and environment to create citizen-first services that improve transportation, way-finding and information delivery. In her session at @ThingsExpo, Laetitia Gazel-Anthoine, CEO of Connecthings, will focus on successful use c...
After a couple of false starts, cloud-based desktop solutions are picking up steam, driven by trends such as BYOD and pervasive high-speed connectivity. In his session at 15th Cloud Expo, Seth Bostock, CEO of IndependenceIT, cut through the hype and the acronyms, and discussed the emergence of full-featured cloud workspaces that do for the desktop what cloud infrastructure did for the server. He also discussed VDI vs DaaS, implementation strategies and evaluation criteria.
In his session at DevOps Summit, Tapabrata Pal, Director of Enterprise Architecture at Capital One, will tell a story about how Capital One has embraced Agile and DevOps Security practices across the Enterprise – driven by Enterprise Architecture; bringing in Development, Operations and Information Security organizations together. Capital Ones DevOpsSec practice is based upon three "pillars" – Shift-Left, Automate Everything, Dashboard Everything. Within about three years, from 100% waterfall, C...
Building low-cost wearable devices can enhance the quality of our lives. In his session at Internet of @ThingsExpo, Sai Yamanoor, Embedded Software Engineer at Altschool, provided an example of putting together a small keychain within a $50 budget that educates the user about the air quality in their surroundings. He also provided examples such as building a wearable device that provides transit or recreational information. He then reviewed the resources available to build wearable devices at ...
Skeuomorphism usually means retaining existing design cues in something new that doesn’t actually need them. However, the concept of skeuomorphism can be thought of as relating more broadly to applying existing patterns to new technologies that, in fact, cry out for new approaches. In his session at DevOps Summit, Gordon Haff, Senior Cloud Strategy Marketing and Evangelism Manager at Red Hat, will discuss why containers should be paired with new architectural practices such as microservices ra...
With the arrival of the Big Data revolution, a data professional is expected to master a broad spectrum of complex domains including data processing, mathematics, programming languages, machine learning techniques, and business knowledge. While this mastery is undoubtedly important, this narrow focus on tool usage has divorced many from the imagination required to solve real-world problems. As the demand for analysis increases, the data science community must transform from tool experts to "data...
Thanks to Docker, it becomes very easy to leverage containers to build, ship, and run any Linux application on any kind of infrastructure. Docker is particularly helpful for microservice architectures because their successful implementation relies on a fast, efficient deployment mechanism – which is precisely one of the features of Docker. Microservice architectures are therefore becoming more popular, and are increasingly seen as an interesting option even for smaller projects, instead of bein...
With major technology companies and startups seriously embracing IoT strategies, now is the perfect time to attend @ThingsExpo in Silicon Valley. Learn what is going on, contribute to the discussions, and ensure that your enterprise is as "IoT-Ready" as it can be! Internet of @ThingsExpo, taking place Nov 3-5, 2015, at the Santa Clara Convention Center in Santa Clara, CA, is co-located with 17th Cloud Expo and will feature technical sessions from a rock star conference faculty and the leading in...
DevOps tends to focus on the relationship between Dev and Ops, putting an emphasis on the ops and application infrastructure. But that’s changing with microservices architectures. In her session at DevOps Summit, Lori MacVittie, Evangelist for F5 Networks, will focus on how microservices are changing the underlying architectures needed to scale, secure and deliver applications based on highly distributed (micro) services and why that means an expansion into “the network” for DevOps.
How do APIs and IoT relate? The answer is not as simple as merely adding an API on top of a dumb device, but rather about understanding the architectural patterns for implementing an IoT fabric. There are typically two or three trends: Exposing the device to a management framework Exposing that management framework to a business centric logic Exposing that business layer and data to end users. This last trend is the IoT stack, which involves a new shift in the separation of what stuff happe...
The emergence of cloud computing and Big Data warrants a greater role for the PMO to successfully manage enterprise transformation driven by these powerful trends. As the adoption of cloud-based services continues to grow, a governance model is needed to orchestrate enterprise cloud implementations and harness the power of Big Data analytics. In his session at Cloud Expo, Mahesh Singh, President of BigData, Inc., discussed how the Enterprise PMO takes center stage not only in developing the app...
The 3rd International @ThingsExpo, co-located with the 16th International Cloud Expo – to be held June 9-11, 2015, at the Javits Center in New York City, NY – is now accepting Hackathon proposals. Hackathon sponsorship benefits include general brand exposure and increasing engagement with the developer ecosystem. At Cloud Expo 2014 Silicon Valley, IBM held the Bluemix Developer Playground on November 5 and ElasticBox held the DevOps Hackathon on November 6. Both events took place on the expo fl...
In his session at DevOps Summit, Tapabrata Pal, Director of Enterprise Architecture at Capital One, will tell a story about how Capital One has embraced Agile and DevOps Security practices across the Enterprise – driven by Enterprise Architecture; bringing in Development, Operations and Information Security organizations together. Capital Ones DevOpsSec practice is based upon three "pillars" – Shift-Left, Automate Everything, Dashboard Everything. Within about three years, from 100% waterfall, C...