 |
 |
In the last decade wireless technologies have evolved in too many directions. One result is a confusing set of semi-meaningless acronyms, e.g. 1xEV-DO and LTE. And, according to Bob Emmerson, all too often the industry puts air interface technology ahead of what really matters — meaningful and affordable end-user services.
I have opinions but don’t think I’m opinionated. I’ve expressed a few in the leadin to this article, but I always check with an expert before publication.
Let’s start with acronyms. LTE stands for Long Term Evolution, but the E refers to the evolution of UMTS (Universal Mobile Telecommunications System). However, LTE was designed from scratch using OFDM (Orthogonal Frequency-Division Multiplexing) technology.
UMTS, which has become synonymous with 3G, was also designed from scratch. In this case the technology was W-CDMA (Wideband Code Division Multiple Access). Therefore LTE is evolutionary in terms of functionality, not technology.
Does this matter? It doesn’t if you work within the industry and know what’s behind these terms, but when they are carried forward into the marketplace they are confusing. GPRS (General Packet Radio Service) was marketed as 2.5G for that reason.
LTE is a hot topic right now and there’s been a lot of media attention about Voice and SMS over this air interface, i.e. the fact that LTE doesn’t handle these services without IMS in the core network. My opinion on this issue used to go like this:
1. LTE was designed for high-speed data, not voice or SMS, and the first LTE devices will be for notebooks, i.e. PC cards and USB dongles. Cell phones will come later — much later.
2. In that case why would operators want to promote voice over LTE when they paid zillions for 3G licenses and their 3G networks? 2G and 3G networks are still delivering robust voice services. Therefore operators are going to milk 2G and 3G for all they’re worth and phasing out these networks won’t begin for many years.
3. LTE rollout will initially be limited to a few early adopters; HSPA has similar spectral efficiency to LTE so there’s no need for many operators to rush; multiformat devices will select the optimum network, i.e. 2G, 3G and eventually 4G will coexist.
That’s what I thought, but to check them out I went to LTE Focus, a conference held in Amsterdam organized by Quadritech, a UK company.
Voice and SMS support in LTE networks
Initially I was skeptical about the need for voice support in LTE for the reasons outlined earlier, but I was surprised to find out that some form of voice and SMS support is almost essential. SMS is a key element of most mobile broadband services, e.g. for configuration, roaming preferred list updates and the new EU-mandated ‘advice of charge’ feature.
Without support a user on an LTE data connection would be unreachable by circuitswitched voice and SMS; that is the real issue, not the need for voice to be carried on LTE. For notebooks with dongles voice is less of an issue, but lack of SMS is still problematic.
Right now there are various incompatible solutions that are relatively tricky to implement.
The 3GPP (3rd Generation Partnership Project) preferred solution for voice in LTE is to use IMS in the core network, which is technically elegant but unlikely to be widely implemented.
Therefore 3GPP has developed an alternative called ‘circuit-switched fallback’ to 2G or 3G. When a voice call or a text message is received, the LTE data connection is abruptly terminated and the connection ‘falls back’ to 2/3G instead. This ‘solution’ is technically inelegant; in fact most users would consider it to be unacceptable since the LTE data connection would be left hanging.
Proprietary VoIP clients like Skype can also be seen as ‘solutions’, but not being able to use the phone’s address book would also be seen as unacceptable by many users. Call quality could be an additional issue.
The fourth solution is to connect Mobile Switching Centers (MSCs) in the 2/3G networks to the LTE network via a gateway. This system is called VoLGA (Voice over LTE Generic Access): I didn’t make that one up. I read about this development in the train on my way up to Amsterdam.
VoLGA is another of those semi-meaningless acronyms but right now it may turn out to be the most practical solution. The VoLGA Forum promotes this approach.
Nokia Siemens Networks (NSN) isn’t a member of the forum. They have yet another solution called Fast Track Voice over LTE. NSN says that the company’s mobile softswitch has Session Initiation Protocol (SIP) signaling capability and the Fast Track software update will allow the mobile softswitch to handle VoIP traffic in LTE networks. That sounds like a better solution than VoLGA; it’s a personal opinion but updating software in legacy hardware would seem to be a better option than installing new hardware.
In a recent Commentary column Andrew Seybold (www.andrewseybold.com) had this to say about the way wireless is going. “… for a while, three generations of technologies (2G, 3G and 4G) will continue to provide wireless voice and data services. Over time 4G will become voice capable.”
In a nutshell, there’s no rush. And if you want to be nerdy LTE isn’t 4G, officially it’s 3G although it is widely referred to as 3.9G. It is only when we get to LTE-Advanced that UMTS officially becomes 4G, but let’s not go there.
Now it gets really interesting
Really interesting because one of the presenters at the LTE Focus event, Moray Rumney from Agilent Technologies, talked about the Implications of the growth in complexity of the wireless ecosystem.
Maybe you’re underwhelmed by the title, but an outsider looking at wireless would find it hard to see the logic of having numerous air interfaces - with unhelpful cryptic acronyms, e.g. GPRS and 1xEV-DO. Does the evolution of 2G to 4G have to involve 2.5G, EDGE, Evolved EDGE, HSPA and HSPA+? Does it have to be so complex?
I was hoping that Moray would cut out the usual presentation noise and give us some meaningful signals and I wasn’t disappointed. His highlights are italicized:
“What would we like? We’d like simplicity and economies of scale. We’d get it by having one worldwide wireless standard; one air interface; one frequency band and one core network.”
Dream on? Moray pointed out that “For 2G, GSM substantially achieved this in 1992. With 3G, UMTS tried to achieve this in 1999 but faltered.” He went on to ask what it would take for LTE to deliver.
LTE’s attributes are impressive and include support for QoS, which would enable robust VoIP. But as he indicated, just because the industry invents a new standard does not mean that success is guaranteed. Cellular technology has made significant progress since the ’70s when bulky, expensive analog handsets appeared, but after 1995 meaningful progress has been patchy and we have ended up with a plethora of air interfaces. Moreover, the interfaces are getting more complex, primarily due to the pursuit of high spectral efficiency and high data rates.
Too Many Failures
Moray listed performance failures that he had experienced in one bad 24-hour period. They included: voice call setup failures; double receipt of an SMS; multiple network-generated SMS spam; MMS sent but received as an SMS (no image); MMS sent but never received; video call failure in one direction but success the other way round.
These failures occur “primarily because none of these end-to-end services falls within the scope of conformance testing. Conformance is falling behind and is no longer sufficient to ensure even basic end-to-end services.”
Smartphones
 |
 |
Bob Emmerson
Bob is a freelance writer living in Holland. He has authored several books and written countless white papers on communications and technology. Bob provides insight in to new applications and how they may impact the channel. See www.electric-words.org