Improving mobile network speeds from an end user perspective


EDITORS NOTE: This article was provided by Flash Networks to demonstrate how their optimization techniques speed up mobile browsing.

Only one-third of mobile Internet users are satisfied with the user experience. Speed of response is the number one complaint. As a result, operators are continuously competing to advertise the fastest download rates and invest significant resources in upgrading their networks. How much of this investment is really reflected in the end user experience?

Available bandwidth vs. actual speed

Research conducted in the UK in June 2009 by broadband benchmarking firm, Epitiro, revealed that the average download rate in high speed networks (HSDPA) in the UK is just under 1 Mbit/s (0.9 Mbit/s, to be exact), only 24% of the advertised download speed and 34% slower than speeds achieved on equivalent ADSL broadband services.

While bandwidth dramatically increases as networks evolve, website download time is not reduced to the same extent. In a comparison of HSDPA 7.2 with HSDPA 3.6, download time was reduced by only 6% (based on lab results from large Asia Pacific mobile operator), demonstrating that bandwidth is not the main bottleneck anymore.

To explain the gap, other factors that influence the effective download rate should be considered. First, unlike feature phones that are limited to mobile-aware sites, smartphones, such as BlackBerry, are used to browse the open Internet. Since open Internet pages were originally built for laptops, Smartphones users enjoy richer content but at the expense of slower download times. In live results from several mobile operators worldwide, Smartphone browsing time for the majority of popular websites was found to exceed 20 seconds, on average, even when measured over high speed networks. Second, Internet access distribution in wide areas in the world contributes to increasing variance in ISP access speeds, and in some cases results in significant additional costs.

However, these two factors cannot entirely explain the discrepancies between bandwidth and actual speed.
The nature of applications over wireless networks must also be considered. Most popular application protocols, such as browsing or downloading, were designed for wire line communication, and are therefore very “chatty”, meaning that many transactions are needed in order to fetch a single page. As every object in the page is fetched separately, wireless network latency becomes a critical factor in the overall download time. Latency over mobile networks is normally at least three times than that of ADSL. In addition, with most applications (FTP, email, file downloads, etc.), fluctuations in bandwidth are misinterpreted as network congestion, causing under-utilization of the available bandwidth.

Because factors affecting speed arise from the interaction between applications and the wireless networks, only optimizers that specifically address this interaction could truly accelerate the user experience. Flash Networks is looking to accelerate this interaction.

Flash Networks Data Optimization

Flash Networks’ Mobile Data Optimization Solution enables operators to provide quicker, smoother, and more reliable mobile Internet browsing and downloading. By applying powerful data optimization techniques based on information about devices and real-time network conditions, Flash Networks’ solution adjusts transmission rates to fully utilize the available bandwidth, eliminates “chattiness”, optimizes content presentation, and more – providing a more consistent, richer, and more satisfying browsing experience.

• To reduce communication chatter during Web page downloads, several techniques are implemented to minimize the number of transactions over the network, increasing the number of parallel communication connections for these transactions. Based on real-time network and bearer conditions, Web page objects are logically grouped to further consolidate browser requests.

• To address sensitivity to bandwidth fluctuation, Flash Networks adjusts transmission rates based on real-time bandwidth availability monitoring, resulting in up to 95% utilization of available bandwidth. Unique to Flash Networks, this adjustment enables download acceleration of up to 30%, even for non-compressible files.

• To avoid repetitive content fetching and Internet access delays Flash Networks stores optimized content on the optimization server for repeat requests per device format and in full compliance with guidelines defined by the application.

Already operating in top-tier networks across the globe, Flash Networks’ solution improves average mobile access speeds by 50% for web browsing and by 30% for downloading of any kind (music, files, emails, etc.). The following results are taken from a large Australian mobile operator, and are based on testing conducted on Smartphones over its live HSDPA 3.6 Mbits/s network.

With Flash Networks’ Web Optimization, Smartphone users are now enjoying up to 50% improvement in web page download time.

Testing was conducted using two phones that were configured with APNs that assure similar routing, except that one route pass through the Data Optimization and one route does not. After selecting representative site list of the most popular sites in the specific country (using sites like Alexa), at least 10 browsing sessions are measured for each site, with the devices running head-to-head to ensure similar networks conditions. The results represent the average result of the browsing sessions.

For more information, see “Data Optimization Benchmark over HSDPA.”

4 Responses to “Improving mobile network speeds from an end user perspective”

  1. 1 jfisher

    Nice sales pitch for Flash Networks disguised as an article…

  2. 2 BlackBerryCool (BlackBerry Cool)

    Improving mobile network speeds from an end user perspective

  3. 3 Yankel

    Interesting timing for this article. Looks like Flash Networks is going through some tough times… Fired 35-45 people today…

    For the Hebrew reading crowd:,7340,L-3331982,00.html,7340,L-3750683,00.html

    And for those who don’t read Hebrew:

  4. 4 iskandar_ahmat (Iskandar Ahmat)

    RT @BlackBerryCool: Improving mobile network speeds from an end user perspective

Leave a Reply