Which Uses Less Traffic: BlackBerry Push Or Microsoft Direct Push?
Which uses less traffic; the BlackBerry or a Pocket PC configured for push mail? There was a time where the BlackBerry reigned supreme and no such questions were asked, ever. However, with the slew of maturing push mail options on the market today, people are now asking. Indeed, at the heart of more than a few forum discussions or flame-wars, and even reflected in many of the search queries that send visitors coming to Technology At Play daily – this is one of the perennial question we encounter whenever there is a comparison of push mail technologies.
The answer is finally answered in a research paper by Rysavy Research. Under sponsorship from Research In Motion (RIM), the folks over at Rysavy Research conducted a series of laboratory tests to measure the data consumption of RIM’s BlackBerry solution versus Microsoft’s Direct Push. This was done in a fully controlled environment simulated by actual handhelds, servers and sample e-mails that mirror real-life usage.
Today, we shall take a look at the findings in the paper and summarize the results as well as conduct a closer examination of what it entails. None withstanding that the research paper was sponsored, we believe that the results would still be relevant to assist us to answer our question – well, at least until the next wave of major upgrades by either RIM or Microsoft.
In the meantime, we will do the crunching of the available facts and figures and explain to you what it all means. But before we do that, let us first take a quick refresher at the importance of achieving high network efficiency in the context of push mail.
Just how important is network efficiency?
As with wired computer networks, it is not hard to imagine that network efficiency is important, especially on a wireless medium such as a mobile network. However, this is especially pertinent especially in the context of a handheld device. Essentially, the reasons can be broken down into:
1. Lower load for network operator
This has to do with the finite availability of capacity in a wireless network. Take a user operating at 1 Mbps downloading a data file. While receiving data, he is in effect consuming one hundred times the bandwidth of a single voice user. Even though data usage tends to be sporadic, the fact remains that the users can consume far more capacity with data communications than with voice communications.
Practically, where a voice user with a 1,000-minutes plan can consume about 100 megabytes (MB) of data in a month, a data user on the other hand can consume the same amount of data in just 10-minutes when downloading a high-resolution video clip. This becomes an issue especially since most operators world-wide are performing a rolling or partial upgrade of their networks instead of a whole-sale replacement. 2G and 2.5G networks with their substantially more finite resources will probably be around for many more years.
However, it must be noted that not many telco offer unlimited data usage plans as yet. As it is, the reality remains that at the moment, the more data you transfer, the more money you are putting into the telco’s coffers. Hence, it might be a moot point at the short to medium term.
2. Battery life
There is a limit to the amount of battery power that a typical handheld will have access to. Bearing in mind that the biggest power consumer on a handheld is its backlit screen and the radio unit, it makes sense that a lower data transmission will translate directly into a much longer battery life.
3. Better user experience
Processor speed and capability has continued to improve in leaps and bounds even in the handheld device arena. However, it is logical that the user experience can still be enhanced by a snappier response where the handheld is less encumbered with the task of processing network traffic.
Without much further ado, let us shift our focus back at the topic on-hand and take a look at how the testing was conducted by Rysavy Research.
On the cellular side, the testing was done with an Agilent 8960 advanced wireless network emulator hardware to emulate an entire cellular operator network in a box. Data services was via GPRS (General Packet Radio Service) – there is no disadvantage here from not using faster networks such as EDGE, UMTS or EV-DO since we are only measuring quantity of data transferred. A shielded transmission box was also used to completely isolate the handheld from any other resident RF signals. To capture the IP data emitted from the Agilent 8960, an Ethereal capture server was used.
For the BlackBerry, a BES (BlackBerry Enterprise Server) was also setup accordingly. There is also a VPN which was used to connect to the RIM NOC.
For the handheld, a BlackBerry 7290 was used as well as a T-Mobile MDA that has been cross-flashed with an i-mate™ ROM. The reason given for the cross-flashing was that a Direct Push-ready ROM was not yet made available for the MDA at the point of the test.
A custom .NET application was used to send off the simulated e-mail messages that are delivered via Microsoft Exchange 2003 SP2 for both setups. The entire test suite of 9 categories of tests was repeated five times for both setups.
As far as we can see, the setup design was robust and as accurately reflecting an actual network as possible.
No related posts.