Hi everyone, I have noticed that there is some misunderstanding about the network sharing between Bell and Telus. Hopefully this will help with some of the confusion. I've worked as a network engineer for operators and vendors and this is an area I am familiar with.
To begin with, let's start with a simple LTE network diagram:
- UE is the phone.
- E-UTRAN is the LTE radio access network, also known as RAN. In LTE, the RAN is only comprised of 1 element, this is the base station, also known as eNodeB. This is basically the "tower" that your phone talks to. In 3G networks, the RAN is made up of NodeB (tower) and RNC (controller).
- EPC is the LTE core network (proper name is Evolved Packet Core), and this equipment is hosted in a few secure facilities across Canada. Data between RAN and core is carried via "backhaul" or transport network (not shown here).
Bell and Telus only share the RAN across Canada. They share 3G and LTE RANs. In some cities (Vancouver), Telus owns the RAN, and Bell users are allowed to use it. In other cities (Toronto), it's the other way around. This allows Bell and Telus to spend less money building out a network, while both benefiting from it. In a city like Toronto, there are only 3 RAN operators: Rogers, Bell, Shaw (Freedom Mobile/Wind). Any other company that sells service is really using the RAN of those 3 companies.
Bell and Telus have independent EPC and IMS networks (IMS is used for VoLTE among other things). The core network is what authenticates user SIM cards, tracks usage and billing, completes calls, delivers text messages, and connects data sessions to the internet. All Bell and Telus eNodeBs in Canada are connected to both Bell and Telus cores.
To make a phone call, or send/receive text, or use data, the whole network must be functional. The RAN cannot deliver any services on its own, and the core cannot directly talk to a phone. So if any of these parts go down, the user will not have service.
If there is ever an outage that only affects Telus subscribers, it's probably due to the core. If there is an outage that affect Bell and Telus subscribers simultaneously, it's probably due to the RAN.
Coverage is 100% due to eNodeB. Therefore, if you have good coverage with Bell or Virigin Mobile, you will have identical coverage with Telus, Koodo, or PM. It can be possible for coverage to be different if Bell and Telus didn't share all bands. However, they currently share all bands except Band 30 (2.3 GHz, which is high frequency and doesn't impact coverage). Band 30 is only for Telus at the moment.
UPDATE
As a result of Bell's 2017 acquisition of MTS, the situation in Manitoba is complicated. In Winnipeg and Brandon, Telus operates a new independent RAN, separate from the shared Bell/Rogers RAN. Coverage between Bell and Telus is different in these markets as a result.
/UPDATE
For a good map of NodeB (3G) and eNodeB (LTE) locations, refer to this link:
https://www.ertyu.org/steven_nikkel/cancellsites.html
After clicking on a site, refer to BW (bandwidth), this may give a clue to the technology being used.
1.25 MHz is CDMA (old 2G network that is not used by modern phones)As of mid-2019, there is no longer any CDMA networks in operation in Canada.- 4.15 MHz is Bell 3G
- 5 MHz can be 3G or LTE
- 10 MHz or higher is LTE
Here is a specific example of how it works in Quebec City:
If you have any questions, I'll do my best to help answer them!
Update: additional LTE topics explained here:
http://productioncommunity.publicmobile.ca/t5/Discussions/LTE-network-fundamentals/td-p/130581
Update 2: key information from this article was used in a 2020 post by Inside Towers, without crediting the original article here.
I'll provide a bit of background before I answer your question.
Wireless networks use spectrum that each operator has licensed from the government to broadcast.
LTE specifications allow flexible channel sizes to be deployed with the following bandwidths: 1.4, 3, 5, 10, 15, or 20 MHz. Wider channels provide faster speeds. The spectrum itself is not related to speeds. For example, 20 MHz of 700 MHz spectrum (Band 12 or 17) would be faster than 10 MHz of 2600 MHz spectrum (Band 7). But nobody owns 20 MHz of 700 MHz, so they usually own wider bandwidths at higher frequencies. That's why people generally say higher frequency spectrum is for capacity, while lower frequency spectrum is for coverage.
A company like Freedom Mobile might have 10 MHz of spectrum in Band 4, so they can deploy a single 10 MHz LTE channel or 2 separate 5 MHz channels (1 for 3G and 1 for LTE). Generally, everyone deploys the widest bandwidth possible at all times.
Example, for Band 4 in Toronto, Bell owns 10 MHz of block F and Telus owns the adjacent 5 MHz of block E.
B4 Block F owners: http://www.ic.gc.ca/eic/site/smt-gst.nsf/eng/sf09011.html
B4 block E owners: http://www.ic.gc.ca/eic/site/smt-gst.nsf/eng/sf09010.html
Since these blocks are contiguous (there is no gap between them), if someone owned both blocks E and F, they would own 15 MHz, which they could deploy as a single 15 MHz LTE channel, even though it's technically 2 blocks of government licensed spectrum. Bell and Telus were given permission by Industry Canada to use each other's Band 4 spectrum, so they own 15 MHz of Band 4 in most of the country, but not everywhere. That's why you don't see a 15 MHz channel for Band 4 in Halifax (Mainland Nova Scotia spectrum tier) on Bell / Telus.
As LTE traffic increases, congestion occurs and more capacity is needed to serve customer demand. There are load-balancing techniques that can be used, but another option is Carrier Aggregation (CA). This allows an operator like Bell or Telus to combine multiple channels into one logical channel.
The specifications allow up to 5 channels to be combined, and as mentioned, each channel can be up to 20 MHz wide. In an ideal scenario, this means someone can deploy 100 MHz CA with 5 carriers.
UPDATE
As of 3GPP Release-13, up to 32 carriers can be aggregated for a total of 640 MHz of spectrum:
http://www.3gpp.org/news-events/3gpp-news/1628-rel13
/UPDATE
In reality, operators don't own so much spectrum, so they still want to combine what they have into CA combinations. For Bell, a good example in Toronto is Band 2 (20 MHz) + Band 4 (15 MHz) + Band 29 (10 MHz). This is 3-CA or 3-carrier CA or tri-band CA, using 45 MHz of aggregated spectrum, resulting in peak download speeds of 335 Mbps.
In technical terms, the main band is called the Primary Component Carrier (PCC), and the additional bands are added as Secondary Component Carriers (SCC). You can have 1 PCC and up to 4 SCCs at one time being used. For Samsung phones, dial *#0011# and you can see the status of CA during a download session. Here you can see B2 PCC and B4 SCC1 and B29 SCC2. Channel sizes are shown as well. In this example, B2 is 15 MHz because this is in Calgary, not Toronto.
One thing to keep in mind is that the total aggregated bandwidth is very important. A single 20 MHz LTE channel (non-CA) will be faster than 3-CA of 5 MHz each (total aggregated bandwidth = 15 MHz).
The eNodeB can configure many different CA combinations very easily, and Bell or Telus can configure 5-CA today if they wanted to, but the problem is that phones are built with very specific CA combinations. Just because a phone supports B2 and B4 in non-CA (regular) mode, it doesn't mean it supports B2 + B4 CA with those specific channel bandwidths. So, you can appreciate why this is a very difficult task for operators. They need to work with phone manufacturers to build devices with very specific CA combinations. This is difficult for Canadian operators, because they are small on the global stage. Samsung and Apple are not going to build customized phones with certain CA capabilities just for Bell or Telus or Rogers. This is one of the reasons why you always see Industry Canada (IC or ISED) follow the USA spectrum regulations (FCC). This gives us a better chance to use phones designed for the USA market, which has a stronger "ecosystem".
To answer your question, CA is entirely a RAN feature, and as long as certain bands are shared by both Bell and Telus (which they are, except B30), then they can all use CA the same way.
Update: Bell / Telus now has quad-band or 4-CA with B2 (20 MHz) + B4 (15 MHz) + B7 (20 MHz) + B7 (20 MHz), for a total of 75 MHz aggregated bandwidth. They have also enabled higher-order modulation, called 256 QAM, which enables total peak download speeds close to 750 Mbps.
Not sure exactly what the high cost is due to. I don't work on the business side, just engineering.
To recap: backhaul is the connection from the eNodeB's baseband unit (BBU) to the core network. There's 2 main types of backhaul connections for modern networks: fiber and microwave.
Bell and Telus have invested heavily in fiber, and most of their sites, especially in urban and suburban areas use fiber directly to eNodeB locations. Rural sites use microwave, and extreme rural sites use satellite backhaul, but that introduces massive latency (delay).
Rogers has fiber as well, but more microwave than Bell or Telus. Part of the reason for this is RAN sharing allows Bell/Telus to focus on building half the country with the same amount of money, so they can deploy fiber more quickly and easily.
Wind Mobile has a lot of microwave due to being a new player. Microwave is much easier and cheaper to deploy, but it doesn't provide the same capacity or reliability as fiber.
A microwave antenna looks like this. You can spot one easily on a site:
As we start to see massive throughput increases with LTE-Advanced and 5G, we will start to see limitations in microwave backhaul and companies like Rogers and Wind will need to catch up.
