At Apis, we get a lot of questions about 5G, and quite a few on 4G, of course. Sometimes we’re even asked about 3G. It’s quite rare these days for people to want to learn 2G, and rarer still 1G, but there is a special question that is not quite that uncommon, and that actually encompasses all Gs. The question is:
“What is a generation, really, and how did we get to number five of them?”
So I figured I could walk you through the history of the five mobile generations, in a quick parade-like manner, so that you can then seem more knowledgeable and pretend you’ve been in the business much longer than you actually have.
Ok, here we go.
1G was the first steps on the path of publicly and commercially available mobile telephony. It was analog in nature, much like a normal radio or old walkie-talkie, and the big difference from just using walkie-talkies is that there was an infrastructure around it. Instead of end point-to-end point communication, the signals were handled in base stations, or cell towers. And the base stations talked to each other via some telephony exchanges so Alice and Bob, with their mobile phones didn’t even need to be communicating over the same base station, but could be in different parts of the country. They could even reach the fixed telephony network (PSTN). I say “different parts of the country” here and not “of the world”, because this was a market of highly isolated mobile networks. There was for instance NMT in Nordic countries, TACS in UK and AMPS in the US. All analog, and the phones were big. Sometimes backpack big, and voice calls was really the only service offered by these networks. The first generation of mobile telephony was first offered to ordinary people around the year 1980.
2G came around 1990 and represented the transition from analog to digital transmission of voice. It improved the audio quality a lot because it was possible to do digital error correction of the bits, and it was also possible to cram more users in on the same amount of radio spectrum, which was of course desirable by operators. The world started to get a little bit more in synch at this point and ITU (the International Telecommunications Union) helped out organizing things so that different parts of the world agreed on radio frequencies to use. Still, there were quite a few 2G technologies. GSM in large parts of Europe and Asia, D-AMPS (a.k.a TDMA) and cdma in the US. Then again, there was overlap so there were GSM operators in the US and cdma operators in Asia as well. But it was not simple for travelling mobile telephoners (that’s what you’d call a person who telephones, right?). It was not uncommon with double or even triple-band phones enabling travelling to different parts of the world, as even if the technology was the same in two different countries, the frequencies in use might not be. A bit of a mess. A cool mess that everybody loved, though, let’s not forget. The 1990s was really the big break-through decade for ubiquitous possession of mobile phones. By the end of the decade, almost everyone had one.
As 2G was digital in its nature, it came naturally that the idea arose to send data (not voice) over the airwaves. The 1990s was also the decade of sudden success of, and public access to, the Internet so mobile data could mean mobile Internet access. It wasn’t mobile broadband, though, let me tell you. It was in the range of tens of kilobits per second. That’s kilobits, yes, you read it correctly.
However, still within 2G, the power to send more data was added. Up until now there hade been only network nodes designed to handle voice calls (telephony switches). Sure, you could use them to put data in there instead of voice, but that was pretty much a workaround. Parallel to these voice switches were added routers. Yes, you heard me, routers. This was the decade of the Internet, after all, and IP was starting to become all the rage. And IP routers seemed to do the job of moving data more efficiently than machines designed for handling voice calls. This was in the GSM context called GPRS, and offered data speeds up to hundreds of kilobits per second. In essence we had one radio access network and two core networks, one for voice calls and one for data transport.
The second generation mobile telephony also saw the birth of texting. A service that originally wasn’t meant for the public at all, but rather for internal signaling and housekeeping, was suddenly exposed to the public, and boy did the public go bananas for this! It only goes to tell you that it’s difficult to predict the killer apps of the future. Even now, with 5G around the corner, good old texting is still used a lot.
So 2G being a bit of a pasta dish when it came to global alignment, ITU stepped in and said “Please get your ducks in a row, people, and make just one third generation mobile telephony system”. ITU called this contest, or order, IMT-2000. IMT for ”International Mobile Telecommunications” and 2000 because 3G was expected to enter stage sometime around the year 2000. They almost got their wish granted.
3G came around 2000 but the world saw the birth of not one, but two, major 3G technologies. Still much better than the multitude of 2G systems, not to mention 1G. The two were cdma2000 (the successor of the 2G cdma system) and UMTS (stemming from GSM and GPRS). UMTS was proposed by 3GPP (third generation partnership project) and cdma2000 by the very cleverly named 3GPP2.
By the way, all these groups, be it 3GPP, ITU or others are mainly collections of people working for companies in the industry who want to affect the way future technology works. They are standardization bodies who write papers that people can use if they want to, and not authorities in any other way. They can be important, if they write a standard that the world happens to really like and everyone wants to use, but they are not in the position to demand anything from any country or company in the world. The 3GPP track is the one that has subsequently taken over the world, and 3GPP2, while having had a horse in the 3G race, did not continue with 4G or 5G. So what we’re seeing today is the evolution of the UMTS created for 3G. UMTS didn’t add anything revolutionary in the core network. It just inherited the two systems from 2G (voice switches and data routers, remember?) and added a super-fast radio access network, enabling user speeds eventually reaching several Mbit per second. Yes, finally Megabits.
The idea – or at least the vision for the future – was to scrap the old way of making voice calls now, and have all telephony be VoIP or Voice-over-IP. 3GPP even designed a subsystem for this, based on the VoIP protocol SIP (Session Initiation Protocol). The subsystem was called IMS (IP Multimedia Subsystem). We had mobile phones with IP access to the network, and nodes in the network that could handle voice calls over IP. And it should have worked. Only it didn’t. It was just too difficult to get the radio connection quality up to par where mobile VoIP calls could compete with traditional non-IP voice. IMS was a really good system for IP telephony though, and was access technology-agnostic, so it found itself (if a technical system can do such a thing) being used for fixed VoIP for a decade. After that decade, we’re now around 2010, IMS finally came into its rightful place – also in the mobile telephony world.
4G was the next contest from ITU. Remember IMT-2000? Well, after that, and having not gotten one 3G system but two, the ITU again said “Well now you’d better dot all your i’s and cross all your t’s because this time we want a really fast mobile system, we’re talking Gigabits per second, and we only want one! Capiche?”. This contest was called “IMT-Advanced”.
And now, finally, only one horse entered the race, and it was the 3GPP horse. It can be debated exactly which 3GPP technology corresponds to 4G. If you ask sales people it’s LTE (the technology believe-it-or-not non-ironically named “Long-Term Evolution”). If you ask standardization nerds, it’s the successor of LTE, called LTE-Advanced, because of course there was immediately a follow-up to the “Long Term”-named technology. (LTE-Advanced also has a successor, called LTE-Advanced Pro.) All the LTE technologies (standardized by 3GPP) are usually considered to be in the 4G family.
LTE changed both radio and core networks with ODFM being used on the radio interface, and only IP-based services supported. It gave users higher bit rates, enabling cool new services over the air. One such service was of course, mobile VoIP. Remember IMS, created for 3G? Well, with 4G it finally worked, and the solution is usually called VoLTE (Voice over LTE, which technically means LTE + IMS).
But ITU haven’t checked out yet. They have more surprises! A few years back they told us about IMT-2020. Just like with IMT-2000, the number is a year, and IMT-2020 signals that the next generation of mobile communications, the fifth, will come around sometime in the neighborhood of 2020.
5G offers new things. The only standard proposing a 5G system comes from 3GPP, is very much a successor of the LTE track, and offers a new radio network as well as a new core network. The radio network is so much an evolution of LTE-Advanced Pro it could almost be called “LTE-Advanced Pro Superduper”, but I guess you have to draw the line somewhere. It is instead now just called “New Radio” or NR (let’s see how long that name feels adequate). The core network is more different than its 4G predecessor, offering among other things support for service exposure to third parties (historically frowned upon in the mobile community) a very flexible way of routing data and a much more Internet-like way of handling things (with HTTP and RESTful APIs, if you’re into those things).
5G offers much higher bit rates, again, which is something the world can never get enough of, seemingly. This is called “enhanced Mobile Broadband” or eMBB in 5G. We’re talking user data rates up to Gigabits per second now. 5G also boasts support for Ultra-Reliable Low-Latency Communication (URLLC) and massive Machine-Type Communication (mMTC). mMTC is 5G lingo for Internet-of-Things (IoT), so that certainly comes in second place after eMBB when it comes to what will be implemented from day 1, while URLLC (created for self-driving cars and remote surgery, for instance) is likely to enter stage a little later. It’s easier to start with the lower-hanging fruit.
5G also places a lot of focus on how the actual infrastructure will be built, emphasizing the need for virtualization, cloud-based systems and software-defined networking (SDN). If you get all of those buzzwords checked off, you’d be able to do another very popular thing launched with 5G, called “Network slicing”. Essentially dividing up a physical infrastructure into separate logical entities working as if they were physically divided. This could be used for dividing services or customers, handling them in different slices, enabling better quality-of-service control.
From 1G to 5G
Well, there you have it. 40 years, from 1980 to now (soon, at least). From analog almost-walkie-talkies to virtualized network slices handling gigabits per second for single users. It’s an interesting tale, with a million cool details left out in this brief article, but it really has a logic to it all the way. It makes sense.