The real ‘G’

Hello tech.geniuses and the rest of the unprivileged world (Sheldon Cooper-like derogation, and thankfully it ends here. You should see ‘the big bang theory’ if you haven’t). I’d like to mention a thing or two about the ‘G’s (geniuses? You ask).

Recently, I was discussing with a friend about how useless our Multilinks Blue® 3G modems which we acquired over a year ago had become, comparing them with those from the GSM operators, he asked ‘aren’t they all 3G?’ (Yes, that G).Well, that depends on how you look at the situation and I hope to explain later. The GSM operators brand their HSPA modems as 3.5G with Etisalat even calling theirs 3.75G, though I haven’t checked out the latest products from the CDMA counterparts (Multilinks, Starcomms, Visafone etc.) Whenever these companies brand their products in such a way, I tend to see it as a marketing strategy, customers would naturally say things like, ‘My 3G modem was pretty fast, 3.5G should be a blast’ or ‘Swift is now providing 4G, awesome!’.

There’s actually a lot more behind the Gs beyond speed (well, speed appears to be the major catch). First of all G stands for Generation, so each G is describing a particular generation of mobile telecommunications technologies and networks with the recommended services they should offer for them to be classified as that G. For example, 3G stands for 3rdgeneration, though I wouldn’t know what to call 3.75G. On a more serious note, the decimalized Gs are actually technologies which are a transition from a former generation to the next, with the features of the next generation not fully realized but well advanced than the former. So I’ll go right ahead and mention a thing or two about the different generations of mobile telecoms networks and technologies.

First Generation: Did you know that there was 1G ?(logically yes, but really?) Ok then, I’ll start with the first generation though there isn’t much to say here. 1G as it is now known emerged in the early 80’s. These cellular networks were designed to basically handle voice traffic only and they transmitted voice traffic in analogue form (talk about inefficient). Capacity was low as they used frequency division multiple access (FDMA) as the multiple access scheme, so once a user was on a frequency channel, no other user could access that channel and we all know how scarce frequency resources are (try using your Wi-Fi or Bluetooth device near a microwave oven). That’s not the even bad news, networks belonging to different companies or countries were not interoperable and coverage area was usually small compared to what we have today so products had limited markets. Examples of 1stgeneration technologies were Advanced Mobile Phone System (AMPS), Total Access Telecommunications System (TACS) and Nordic Mobile Telephone (NMT). I bet you’ve probably never heard of these technologies and neither had I. 1G wasn’t totally useless as it paved the way for the newer generations to come in and people generally became aware of the potentials in mobile communications.

Second Generation:I’m usually disgusted whenever I think of how long it took Nigeria to catch on with 2G considering the fact that it actually started off in the late 80’s/early 90’s (we all remember GSM started here just 10 years ago, the good thing is that we aren’t too far behind now…well, somewhat). With the second generation, the basic problems with 1G were identified and tackled head on. Initially, it was proposed to handle voice traffic, SMS and fax, but with the need for data services becoming apparent enhancements were made. One great thing about 2G is that it transmits information in digital form and this allowed for better voice transmission, greater security through encryption and easy integration with other digital systems like ISDN and the internet. GSM turned out to be the most successful 2G standard with its support for international roaming. GSM increases capacity by combining FDMA with TDMA (time division multiple access) where each frequency channel is divided into timeslots and each user allotted a timeslot which periodically repeats itself in a very short time (about 577µs). A maximum of 8 users per frequency channel can be supported (for full rate, though two users can be made to share a timeslot under half rate) and the number of channels would depend on the number of transceiver modules installed. We know the GSM operators – MTN, Glo, Airtel, Etisalat in Nigeria and others like Vodafone, T-Mobile and AT&T abroad.

Another popular 2G standard is CDMA (code division multiple access), commonly known as cdmaOne or IS-95 which is much more popular in North America. It turns out CDMA was originally developed by the military (its invention I think was credited to a female spy who doubled as an actress). CDMA carries out channelization by using different codes to differentiate users who are all on the same frequency. The CDMA operators in Nigeria include Starcomms, Visafone, O-net, Zoom etc.

In order to understand how the multiple access schemes work in 2G technologies, think several people in the same room speaking different languages with each pair at a low tone so as not to interfere with each others in CDMA, while in GSM users are compartmentalized into smaller sound-proof rooms so they don’t interfere.

Second Generation Evolution: Typically known as 2.5G, the need for support for data services was realized and traditional GSM could only support data rates of 9.6 kbps (full rate) or 14.4 kbps (half rate). Three technologies emanated during this era as far as GSM is concerned, two of which we know well. The earliest was High Speed Circuit-Switched Data (HSCSD). In HSCSD, more timeslots were combined (maximum of four per user) to give data rates up to 28.8 kbps uplink and downlink (2 timeslots each) or asymmetrically 43.2 kbps downlink and 14.4 kbps uplink (3 to 1).

The next was General Packet Radio Service (GPRS). GPRS is packet switched so users will always get an ‘always on’ service characteristic of the internet. A lot more was changed to implement GPRS, new logical channels were defined along with new coding schemes. Also new elements that connect the existing GSM core network to the IP-Internet network were introduced. The coding scheme used and multiple timeslot allocation can allow data rates up to 171.2 kbps (theoretically).

The last 2.5G technology (which is referred to as part of 3G in some texts) is Enhanced Data rates for GSM Evolution (EDGE). EDGE further improves data rates by using a more advanced modulation scheme known as 8 phase-shift keying (8-PSK) to a maximum of 48 kbps per timeslot (i.e. 384 kbps per channel -8 timeslots). It involved some upgrade between interfaces connecting the Base station to the Base Station controller as well as installation of new EDGE transceiver.

We’ve all seen GPRS/EDGE capable phones and frankly they are the most common technologies used to access the mobile internet in Nigeria with 3G coverage still limited.

So with 2.5G in place, the stage was set for 3G to come on board.

Third Generation: the third generation technologies were designed to support multimedia communications – pictures, sound, and video. Hence large data rates that would sufficiently support these services would be required. Different standards would be developed but in order for them to be called 3G, the International telecommunications Union (ITU) created the IMT-2000 standard which would define characteristics of any 3G standard some of which are:

· Data rates up to 2 Mbps (depending on mobility)

· Worldwide usability.

· Support for Packed switched and Circuit switched services (typically data and voice)

· High spectrum efficiency, security, quality of service and reliability.

The three most recognised standards are UMTS (Universal Mobile Telecommunications System, which uses Wideband-CDMA and evolved to the HSPA we know today) which is an evolution of GSM (2G) to GPRS/EDGE (2.5G) to WCDMA (3G), CDMA2000 which is an evolution of the original CDMA (IS-95) which is where our Multilinks Blue 3G modem lies and Time Division Synchronous CDMA (TD-SCDMA) which was developed by the Chinese and frankly I know pretty much nothing about at the time of this write up.

I’ll focus more on UMTS/WCDMA from here on which is what is used by the GSM operators in Nigeria for their 3G upwards networks. WCDMA which is the radio access scheme used under UMTS is similar to the original CDMA except the frequency band is much wider hence WCDMA (A carrier band is 5MHz in WCDMA and 1.25MHz in CDMA2000). This allows for higher data rates with WCDMA due to a higher chip rate (3.84 Mcps for WCDMA and 1.288Mcps for CDMA, never mind that for now).

UMTS/WCMA has gone through different evolutionary stages, the original known as 3GPP Rel99 (3GPP stands for Third Generation Partnership Project, a group of standardizations organisation from different countries working on UMTS as the 3G technology, while Rel99 is the December 1999 release for UMTS) involved a change in the radio access method of existing GSM networks (from TDMA/FDMA to WCDMA) and the introduction of elements necessary to support the change like new antennas, base stations (now known as Node Bs), new radio network controller (RNC) with very little changes to the existing GSM core network.]

Remember the earlier 3G phones like the Sony Ericsson k800i or z750a? Connecting these to your PC would give a typical Rel99 connection.

sony ericsson z750a

Third Generation Evolution: Further enhancements where made to the earlier release of UMTS. In 3GPP Rel4, time division duplex was introduced to WCDMA and more changes were made to the core network to further improve data rates. 3.5G is typically release 5 which introduced High Speed Packet Access (HSPA, from which we have HSDPA – downlink and HSUPA – uplink). In HSDPA, data rate is increased to as much as 14 Mbps with increased capacity and reduced delay, particularly for file downloads. This was achieved by introducing a new channel known as HS-DSCH (high speed downlink shared channel) with a more efficient modulation scheme known as 16-Quadrature amplitude modulation (16-QAM, earlier WCDMA used quadrature phase-shift keying – QPSK). Also a mechanism known as ‘fast scheduling’ is used to give preference to users with better radio conditions (better signal strength, less interference etc.) so those users will have better experiences than others under poor radio conditions rather than a round-robin mechanism where every user is given equal priority irrespective of the quality of their radio environment. The point is that users under good radio conditions can use network resources more efficiently, unlike WCDMA where power control is used to increase transmission power to users under poor radio conditions. Think of HSDPA as the real ‘downloads monster’ (well, for now and for cellular networks).

The modems being offered by GSM operators in Nigeria today (MTN Fast link, Etisalat Easyblaze, Glo NetPro etc.) are typically HSPA capable and are backwards compatible with WCDMA and EDGE depending on coverage and radio conditions.

3gmodem mtn fastlink huawei modem

Now, I’ll cautiously say 4G really isn’t being implemented here in Nigeria yet though other countries are just beginning to roll out commercial networks tagged as 4G and typically Nigeria may follow suit like 2-3 years after (better than the 10+ years it took to deploy 2G, though I think CDMA might have been in Nigeria slightly before then). You might ask, how about Swift’s or Mobitel’s 4G? I’ll get to that soon. Ok, on to pre-4G and 4G then.

Fourth Generation:ITU came around with their specifications for the fourth generation of mobile networks, this time tagged IMT-Advanced and some of the features were:

· These networks would be fully IP based (voice communications would no longer be on circuit switched networks)

· High quality of service for multimedia applications

· Data rate up to 1Gbps (mad or what?) or 100Mbps for fast moving nodes

· Dynamically shared network resources between users.

Two candidates emerged as the 4G technologies namely of choice, Long Term Evolution (LTE) Advanced from the 3GPP (yes, the GSM to GPRS/EDGE to UMTS to HSPA people again) and IEEE 802.16m or WirelessMAN-Advanced from IEEE.

Prior to these two, there were the original LTE and Mobile-WiMAX (IEEE802.16e) both from the same guys as above. They are often described as 3.9G technologies as they do not fully meet all the IMT-Advanced requirements but were branded as 4G by those who rolled them out commercially.

Now, to the Swift and Mobitel 4G which I must say is not really 4G (definitely not at all!!!). Firstly, the ‘G’s as I mentioned earlier describe mobile cellular networks, Swift’s isn’t by any means mobile (and don’t say the modems are portable) and secondly, I believe they offer WiMAX (not even mobile-WiMAX, though I stand to be corrected) and WiMAX is certainly not a 4G technology. It’s almost like saying WiFi or 802.11n is 4G. As I mentioned earlier these are just marketing strategies. Anyway, 4G is supposed to be badass and all; I hope it comes around soon enough and I came across a document about 5G a while back but didn’t bother to read through. To summarize, I’ll use a diagram which I obtained from the book “UMTS Networks Architecture Mobility and Services, Wiley 2005”.

Evolution of Mobile Telecoms Networks

I hope this write up was a little enlightening and hopefully will help not to mix up terms though I barely scratched the surface on the individual technologies (not that my fingers are deep enough, I’m just learning myself). I’ll be glad to receive any corrections and pointers and I’m sorry for not citing my references within the article, it’s a really bad writing habit.

Artificial Intelligence and the Future of the Web

Artificial Intelligence and the Future of the Web

Artifical Intelligence

I was watching a Ted talk by Tim Berners Lee and he was talking about the semantic web. Its not by any means new to anyone on in Tech. But as he was talking I got a revelation from it. It was like I caught his passion even though the audience did not see anything to get excited about (just like I did before).

Let me give you some detail and I hope that you also get the idea. Tim who invented the world wide web said that the idea of the World Wide Web was out of frustration. Frustration in the sense that he was exposed to hundreds of scientific materials of great value but they were in all kinds of formats and required all kinds of proprietary software to view them. This made integrating and combining information contained in the documents very difficult if not impossible. He invented the www so that documents could be in one format and so that it would be easy to ‘connect’ documents together.

Today, we have benefited immensely from this vision of an open www where documents are linked in a fashion that knowledge is not only accessible but easier to connect and now, we have another dilemma. The structure/order of data. Everything that we do on our computers is all about data. This data is in different formats and stored in different places which makes gathering and harnessing this data a real pain.

Solution: Semantic Web – In a semantic web, it would contain just raw meaningful data that can be linked to other raw meaningful data. Taking this data and making a web out of this data is going to change the world. Take a look at wikipedia. Even though its just web pages, the way the data is linked and connected produces very powerful effects. Like say I pick an article of a person. And it tells me that the person attended an event which would be a hyper link to the event and then reading about that event I get a link to its location say Germany and the I go into Germany and then I can learn about the country, its population, heritage, culture e.t.c That’s what a semantic web would look like

Now if in the midst of this chaos of documents which is the world wide web, Search engines like Google are able to quickly parse through and harness useful data as to be able to even translate text from different languages and guess what you are about to type before you even type it, imagine if devices and software had access to a web of data instead and they had clever ways of parsing through all that data…. The results would be pure Artificial Intelligence. The kind that would be astonishing. Imagine a world where as you walk past a restaurant, your phone reminds you that you have been to that restaurant before, tells you about their favorite dish and how its prepared. Imagine a world where computers could make complex reports based on a mash-up of data from from millions of sources… This… my dear friends is the future.

Blog Upgrade

Wow, This isn’t the article I thought I’d be writing. But it looks like I will not be able to keep up with blog posts as I am now twice as busy as I was before. That’s not to say that I’m abandoning this blog, Of-course Not! Its just that I am about to create my own start-up (Trust Me it is HUGE!). Its a really challenging quest that I have taken up and so I will not be able to post quality content in the mean time. However, you can still follow me on Google+ to get Tech Gist and other random stuff.

Blog Upgrade

One of the reason for this decision is that I have come to realize some of the limitations inherent in WordPress hosting service (Props to them for making it free). An example is that I cannot install plugins which means that I can’t display blocks of code in a neat format which is only available using plugins. So I decided to move the blog to a proper hosting service and custom domain name.

Multi-Media Content

In addition to the  blog upgrade that’s coming soon, We will now add more multi-media content like videos and audio/slides. So you can look forward to that. I know we have serious bandwidth issues in Nigeria but things are changing. Wayne Gretzsky said “I skate to where the puck is going to be, not where it is”

Co-Authors

This is not to say you should not expect any posts from this blog as some other cool authors like Sadiq Mohd, Eze Amechi or John Obidi could post at any time. Its just that I can no longer keep up. 😦

As sad as it is, I know its for the best. I believe that this timeout is necessary to come out stronger. So see you soon