Active Ethernet vs. GPON

[pinball wizard]

It is actually not the cost of putting fibre in the ground that is that much different from player to player. It is the time over which they want to make a return on investment. The smaller and more unknown players have long ROI periods (10-12 years), while the Vodacoms and MTNs have a very short ROI period (2-3 years). This is similar to taking out a loan and changing the time over which you have to repay that loan. The shorter the time, the larger the installments. Sure there are some differences in the cost of rolling out fibre from provider to provider but this is not close to the largest part of the equation.

So what you mean is that MTN and Vodacom could be selling this service much much cheaper than they do?

 

[DERoestorf]

So what you mean is that MTN and Vodacom could be selling this service much much cheaper than they do?

Yes, they can. However they prefer not to. What they do is install Fibre. Pay it off as fast as possible (at the expense of their customers) and once it is paid off, they coin the profit, never reducing the price. Essentially screwing over their customers all the way to the bank.

 

[pinball wizard]

Yes, they can. However they prefer not to. What they do is install Fibre. Pay it off as fast as possible (at the expense of their customers) and once it is paid off, they coin the profit, never reducing the price. Essentially screwing over their customers all the way to the bank.

Oh, you mean like they do with sms rates, call rates, mobile data prices? Business as usual.

 

[DERoestorf]

Oh, you mean like they do with sms rates, call rates, mobile data prices? Business as usual.

Exactly like that.

And then when people don't use overpriced and overcharged sms's and start using whatsapp, they start crying and threaten the consumers that they will start charging us for using it.

Ridiculous.

 

[Chris.Geerdts]

GPON solutions result in the users being 'locked in' to GPON. So a move to XG-PON or NXG-PON would be relatively expensive and might require trenching. Who will pay ?

I am interested in this 'finding'. The whole point of GPON evolution is to re-use the fibre by adding lambdas and some kit on the CO side.

I remember having the GPON vs Active Ethernet debate a while back and since then literally millions of households have been connected globally across all the continents and it continues to have traction. Many of these countries are on much higher average throughputs than South Africa, where people still consider 5 and 10 megs to be broadband. All access or last-mile speeds globally are going through exponential increases, but South Africa lags behind dozens of other countries (as we could see when Ookla used to publish the charts).

Vumatel was absolutely clear from the outset that GPON was inferior and they would use active ethernet, yet here we are today .... I understand that the switch was linked to their buying Fibrehoods, but I think there is more to it than that.

MTN, Vodacom and Telkom, who are all highly skilled in deploying active links, also use GPON for FTTH.

I'm not personally invested in this either way but am curious to understand the positions on this. At the end of the day, workable economics is not only about making profit but about speeding up fibre deployment to those who don't yet have.

 

[cobusv]

Vumatel was absolutely clear from the outset that GPON was inferior and they would use active ethernet,

This is not true, there is absolutely no difference in using either Active ethernet vs GPON. Gpon will usually get deployed in areas that has a higher density and are perfectly suited for MDU's whereas active ethernet will be rolled out in areas where there are more SDU's. but either will work.

At the end of the day the user experience will be exactly the same weather you in a GPON area or a Active Ethernet area

 

[cobusv]

The whole point of GPON evolution is to re-use the fibre by adding lambdas and some kit on the CO side

nope, not correct, it does not use lamba's but rather use splitters withing in the network, the rollout is much faster and easily scalable, whereas with active ethernet it's point to point fibre and gpon uses the concept of 'splitting'. you can think of it as using TDM technology.

 

[ThatOtherDude]

GPON is cheaper to roll out
less fibre in the ground & less splices.

Active Ethernet is faster/easier to troubleshoot as network errors are automatically sent to the NOC from the switch.
With GPON a splice team needs to go out to investigate where the issue lies.

AE requires significantly more rack space.
GPON supports Tripple Play out the box.
AE doesn't. (requires WDM upgrades)

End users normally dont know /care what they're using.

If I had 100Million doleros I'd go with AE every time.
If I had 100 doleros I'd go with GPON (whiel telling everyone how great it is)

 

[system32]

Lots of talk about speed in this discussion - which is great for streaming.

Is there a latency difference between AE and GPON?
Which is better for gaming?

 

[Jaseyyxx]

Lots of talk about speed in this discussion - which is great for streaming.

Is there a latency difference between AE and GPON?
Which is better for gaming?

Honestly no difference in latency

 

[InTheCube]

It is now 2022, 7 years since the start of this discussion. Where does the market stand with regards to this debate? Who is using what? What are your guys updated opinions?

I am on an Openserve 500mbps/250mbps fibre line through Afrihost (the fastest line speed they offer). I am almost 95% certain that they rolled out GPON in our suburb (Lenasia). I get my full line speed ALWAYS, irrespective of time of day, peak demand, etc.. The same applies to my neighbours, friends, relatives and colleagues who also live in Lenasia and are using Openserve fibre. Everyone is super happy. No complaints. Stable as a rock (minus 3 major events in the last 5 years lasting less than a day of downtime/slow speeds relating to backhaul vandalism and/or equipment failure at the CO). If you have your own generator / inverter / battery backup, load-shedding and even larger/longer power outages doesn't affect internet connectivity.

Openserve has been performing marvelously from everything I've heard and experienced. Apparently Openserve's rollout of FTTH in Lenasia has been one of their biggest success stories in the whole country. Uptake is massive.

 

[Jaseyyxx]

Yeah openserve always rolls out GPON, it's really only vuma who do AE depending where they roll out

 

[Johnatan56]

It is now 2022, 7 years since the start of this discussion. Where does the market stand with regards to this debate? Who is using what? What are your guys updated opinions?

Still the same opinion as in the beginning, GPON is fine, you have upgrade paths like XGPON (G.987) that have been a standard since 2010, so you knew you could upgrade if the demand was there way before anyone rolled out.
2017 reference costs from: https://www.sciencedirect.com/science/article/pii/S1573427716300522



Fig. 3. Cost of multiple PON deployments for GPON, XGPON, NGPON2 and UDWDM PON in the six bit-rate-demand scenarios specified in Table 9.

So basically once enough people want to go 100Mbps+ we'll start seeing XGPON upgrade deployments, you can also upgrade just parts of the network that need it. Costs of equipment has also dropped quite a bit in the last couple of years.

 

[Kyoto]

Still the same opinion as in the beginning, GPON is fine, you have upgrade paths like XGPON (G.987) that have been a standard since 2010, so you knew you could upgrade if the demand was there way before anyone rolled out.
2017 reference costs from: https://www.sciencedirect.com/science/article/pii/S1573427716300522
View attachment 1362857
View attachment 1362859
View attachment 1362861
Fig. 3. Cost of multiple PON deployments for GPON, XGPON, NGPON2 and UDWDM PON in the six bit-rate-demand scenarios specified in Table 9.

So basically once enough people want to go 100Mbps+ we'll start seeing XGPON upgrade deployments, you can also upgrade just parts of the network that need it. Costs of equipment has also dropped quite a bit in the last couple of years.

Actually XGPON has already been surpassed by XGSPON, unlike XGPON, XGSPON is 10G symmetrical, and is now being deployed in multiple UK and US networks, offering residential speeds of upto 2.5Gb/s. 25G PON and 50G PON coming.

 

[Johnatan56]

Actually XGPON has already been surpassed by XGSPON, unlike XGPON, XGSPON is 10G symmetrical, and is now being deployed in multiple UK and US networks, offering residential speeds of upto 2.5Gb/s. 25G PON and 50G PON coming.

Sure, I'm saying when GPON was rolled out that an upgrade path already existed, not that it's the only/best one anymore.

 

[Chris.Geerdts]

Yeah openserve always rolls out GPON, it's really only vuma who do AE depending where they roll out

Some of us were correct 7 years ago - all the fibre operators use GPON for households now and some fibre operators for business two. You see active ethernet for selected corporates and some (not all) office parks or buildings.

 

[Chris.Geerdts]

Sure, I'm saying when GPON was rolled out that an upgrade path already existed, not that it's the only/best one anymore.

Most ISPs now seem to offer gigabit service but Plum Consulting, at an FTTH Council Europe sponsored by Nokia was saying recently that only 15-20% of subs selected it.

It is early days for 25G PON but SA still hasn't seen much need for 10XS. Most households buy entry products.

 

[Vorastra]

Most ISPs now seem to offer gigabit service but Nokia was saying recently that only 15-20% of subs selected it.

It is early days for 25G PON but SA still hasn't seen much need for 10XS. Most households buy entry products.

How does Nokia know the speed uptake for FNOs' customer base?
Maybe I'm misunderstanding who you're referring to as "subs".
If we're talking about an FNO's overall customers, 15% uptake of 1Gb speeds sounds surprisingly high to me.

 

[Johnatan56]

Most ISPs now seem to offer gigabit service but Nokia was saying recently that only 15-20% of subs selected it.

It is early days for 25G PON but SA still hasn't seen much need for 10XS. Most households buy entry products.

I am highly doubtful of 1:5 taking a 1Gbps product if it's the end consumer, otherwise I don't understand the rest of your message, missing context. Do you mean gigabit ports or something at the switch it connects to?
Why not is mostly price, consumers will generally buy at a price point rather than speed, and 1Gbps on e.g. Frogfoot is around the R1600pm marker, considering the average income in South Africa, I am highly doubtful anyone outside of either working from home in the media space or extreme enthusiasts would get that, think the cap for most enthusiasts/larger households is the R1k marker, so 100-200Mbps depending, and the norm will be the <= R500 packages, so <=50Mbps packages when possible.

 

[Polymathic]

How does Nokia know the speed uptake for FNOs' customer base?
Maybe I'm misunderstanding who you're referring to as "subs".
If we're talking about an FNO's overall customers, 15% uptake of 1Gb speeds sounds surprisingly high to me.

I am highly doubtful of 1:5 taking a 1Gbps product if it's the end consumer, otherwise I don't understand the rest of your message, missing context. Do you mean gigabit ports or something at the switch it connects to?
Why not is mostly price, consumers will generally buy at a price point rather than speed, and 1Gbps on e.g. Frogfoot is around the R1600pm marker, considering the average income in South Africa, I am highly doubtful anyone outside of either working from home in the media space or extreme enthusiasts would get that, think the cap for most enthusiasts/larger households is the R1k marker, so 100-200Mbps depending, and the norm will be the <= R500 packages, so <=50Mbps packages when possible.

I seem to remember an article saying that Vuma Core uptake in an area was around 20% of households.

So I would assume that 15% uptake refers to fibre customers in an area and not total households