Steps towards better mobile user experience (March 24, 2008)

I’ve stumbled upon the Mobile Wireless Server Developer vacancy in Google UK and asked yourself a question: what would I personally do as a server software developer to deliver a better user experience of rich web applications on mobile devices? It looks like the problem could be split in several parts:

• Device computing capabilities. Rich web applications (read “applications using lots of JavaScript”) require the device to be powerful enough. This is the simplest of all parts because actually, there are devices powerful enough already. There are at least two bold players in this area: iPhone/Safari and Nokia S60 platform’s browser, and this number will inevitably continue to grow, thanks to Moore’s law. However, if this trend would not exist, as a server developer I would implement a front-end proxy a) striping out JavaScript which was tagged by web application developers as “non-mobile” (say, drag’n’drop functionality of Netvibes), and b) executing as much of the remaining JavaScript at the proxy’s side, resulting in the page which look is closer to final render (say, preloading and arranging mashed contents in advance). At this point we go to the next problem:
• Narrower bandwidth. Several possibilities exist for a server developer. Content could go through proxy doing markup conversion and compression like this is done in Opera Mini. Applications could be made bandwidth-aware to queue data transmission when possible.
• Device rendering capabilities. Static media (pictures), by the way, could also be processed by the proxy or the server itself (reduced in size). Critical operation of zooming out JPEGs could be performed quite fast using low-frequency selection from Discrete Cosine Transform curves that form the JPEG image. I was unable to find any information about Opera Mini’s proxy doing this or not.

Telecom liberalization: now with femtocells, postponed (March 8, 2008)

While December 2007 have passed with a question in the air on “will 2008 be a year of the femtocell?”, it is still too early to say about the market disruption for MOs and NEPs. Telco 2.0 reminds us in this context of the 2006 partial acquisition of Spring Mobil, a Swedish innovative femtocell production company, by Tele2. This definitely will be the MOs strategy in the first time of the femtocell existence. By the way, this is what Cisco does to any promising network-related Silicon Valley’s startup: acquires it.

However, telecom market liberalization (and this is exactly what happens when a company may enter a MVNO world for mere €10‒20K) which will happen in the real year of femtocell (this will be no earlier than 2009) will rise a much wider need in the standardized service execution environment. Open source SLEEs (think of Mobicents), as well as commercial ones promise to be in the high demand in this situation.

And while “Telco 2.0” has already become a common buzzword, this situation with the market liberalization which now only promises to appear will be for telecom world the point when it will start to accumulate products and services quantity. It will take several more years for this quantity turn into quality â€” just in the same way, as quantity of web application has lead to a qualitative change in the Internet world (we all here know the buzzword for this change).

However, the in this case the future seems even more bright than before.

Telecom liberalization: a long-awaited change (February 13, 2008)

When I was working in a software development company dealing with JEE and aiming for custom enterprise applications, there were always a tendency to a slight technological change from one project to another. Gurus were hanging out in the smoking corner with a magic words on their lips: “Spring! Generics! Tapestry! Ruby on Rails!” A current favorite technology of a decision-maker in a new project was lobbied at its beginning, added or removed during the project evolution or vetoed at any point. I can’t remember project failures influenced by magic word choice: after all, Spring was just a frameworks and every B.Sc. owner could use generics.

Telecom world is evolving much slower. Projects tend to be longer, key players in this market are rarely changing, as well as technologies around. However, there’s still a room for a slight change: if telco mashups like the award-winning After Hours Doctor’s Office would be as fast to implement as web ones, telco landscape might change for better. A software company could roll-out this solution in one bundle, including a contract with SIP provider (for VoIP termination of PSTN customer calls), and any SIP-based service delivery platform with the corresponding application installed. Applications like mentioned one could be productized, and the more lightweight the solution is, the more opportunities there are for the vendor because of the shorter time to market.

Me personally would be happy to see another bunch of telecom gurus in the smoking corner discussing the magic words offered by the new SDP technology. And you?

IMS and WiMAX (February 5, 2008)

It is clear that despite all the delays in the standartization and implementation procedures IMS will be starting its deployment sooner or later â€” probably as a service layer of 3GPP Long Term Evolution. This set of specifications, apart from the fact that it is purely IP-based, is especially interesting because it does not limit the way how terminal equipment should access the base station and, after all, the core network.

Among the air interfaces (or access modes), WiMAX looks better than others (actually, there’s a “battle” between LTE and WiMAX which I found overhyped, and therefore will write about this some other day). It is defined as a set of IEEE standards (802.16), and there are some chipset implementations of it. These chipsets will inevitably enter the device market quite soon and this probably will be considered as a good opportunity by the mobile operators willing to seamlessly enter 4G/LTE world. As LTE is not yet finished set of standards now, deploying IMS before 4G All IP Network will be in place, and using WiMAX as an access mode, may help operators to test their environment and then switch to AIPN more smoothly. At least, all the necessary technological background is in place for that.

WiMAX fits well in IMS infrastructure for several reasons. First of all, it is fast enough. Depending on the distance from the base station speeds up to 70 Mbits/s could be delivered, though the most common speeds in the city probably would be around 10 Mbits/s, which is almost as good as best-case HSDPA downlink speed right now. Mobile WiMAX (802.16e) offers mobility between base stations. Quality of service is guaranteed for each connection independently.

It looks like next years will show, whether mobile operators will use this opportunity, and would WiMAX be a “helping hand” for IMS deployment all over the world.

Wired on iPhone changing mobile industry; Cringely on Moore’s law (January 27, 2008)

WirelessMoves’ Martin Sauter has posted a link to the interesting article in Wired. It describes the shift in the trends of the mobile world that we are witnessing now: focus of customer attention is moving from operators to device manufacturers. What is even more interesting, that this article in a mainstream technical magazine may probably serve as a milestone marking the point when this trend becomes obvious not only for industry insiders and thus completely inevitable.

Robert X. Cringely predicts that Moore’s law will stay here for at least 15 more years. I don’t want to make any statements about the prediction itself (better read the article), but one related thing comes to my head immediately: we have HSPA family support in our mobile phones thanks to the performance doubling. UMTS infrastructure remained the same, however, user terminals are now capable of performing HARQ — hence speed grows.

← Later entries | Earlier entries →