Incidentally, this also applies to the energy distribution (electricity) of buildings/districts.

It’s a shame there are no flat rate tariffs there LOL

---

@StefanSch wrote:

to shit tüpfli:

Strictly speaking there are always 2 splitters:

1:2 and then 1:32 or

1:4 and then 1:16

Has the advantage that you have to split it into many individual fibers very close to the customer (sleeve or BEP).

---

I believe the variant you are describing corresponds exactly to the construction method that was banned by the Comco about 2 years ago 🙂

Currently, optical splitters are no longer permitted decentrally on the last mile to the customer, but only in the headquarters (CO) 🙂

---

@StefanSch wrote:

to shit tüpfli:

Strictly speaking there are always 2 splitters:

1:2 and then 1:32 or

1:4 and then 1:16

Has the advantage that you have to split it into many individual fibers very close to the customer (sleeve or BEP).

---

Don’t get upset 😁 1:4 and then 1:16 are in P2P in front of the OMDF.

@StefanSch @millernet

OLT with splitter to OMDF

Splinters in the OMDF

https://www.ccm.ch/de/item/caro-splitterkassette-8mm-plc-953351

Splitter cassette in BEP

@user10 wrote:

Splitter cassette in BEP

---

We no longer build splitters into the BEP… and existing ones are replaced with normal splice cassettes for P2P.

Thanks for the pictures @user109. So according to the picture

*OTO
*FD

• BEP
  *Drop
• Street cabinet / manhole
• Feeders
• POP / “Central”
• Splitter OMDF: 16
• Splitter before OLT: 4
• OLT Port: 16 ports per line card
• OLT line card: 7 line cards per OLT (installed 3)
• “Uplink” OLT: 2 × 10 Gbit/s (I’m assuming a LAG now… although the LED on an uplink doesn’t light up at all.)
• Aggregation OLT per POP:?
• Connection to core network (routing etc.):?

This results in an overbooking between OTO and aggregation in the core network of:

• According to the picture with 3 line cards and 2 × 10 Gbit/s uplink: 16 × 4 × 16 × 3 × ½ = 1:1536
• with 7 line cards and 2 × 10 Gbit/s uplink: 16 × 4 × 16 × 7 × ½ = 1:3584
• Full expansion with 7 line cards and 8 × 10 Gbit/s uplink: 16 × 4 × 16 × 7 × ⅛ = 1:896

So: Depending on the configuration, several hundred to several thousand customers 1 × 10 Gbit/s or approx. 8.8 Gbit/s net 😁.

And yes, in practice this will work:

• Simultaneity factor
• Only a few customers are speed test junkies and even know how fast their Anschluss is.
  *Not all customers have 10 Gbit/s as a service profile.
• Customers with 2.5 or even 10 Gbit/s equipment or Network are also rather rare.
• Most devices are connected via WiFi. In practice, gigabit is already the highest feeling there.
• Everyday applications, even large game downloads from XBox, Playstation or Steam etc. reach somewhere at 1 Gbit/s. With XBox and Playstation you’re more likely to get less than that and if you’re lucky you’ll get maybe 500 - 800 Mbit/s.
• The “bandwidth” or Data transfer rate is only one quality criterion of an “Internet connection”. Latency, jitter and packet loss rates are much more important.

Basically these are just the old telephone exchanges. But there was/is a program to reduce the number of centers in order to reduce costs. But I don’t know how far this has progressed.

@EindeutigerUserName the rule is that Swisscom uses 10 km modules, which is usually sufficient since the control centers are a maximum of 6 km away from the customer.

@5018 There are such efforts in Basel, for example to close down the main post office on Freienstrasse and move the whole thing to Wallstrasse. With Fiber, location no longer plays as big a role as it did with Copper.

I’ve asked myself that too. Now with the (involuntary) change in strategy towards P2P, the headquarters respectively. POP’s have a “central” meaning to:

• Every household receives a P2P connection with 1 - 2 fibers directly from the central office, comparable to POTS. Every household or The usage unit once had a two- or even four-wire line to the headquarters.
• Many of these fibers (e.g. feeders) now need to be tightened. In my case, the entire community would have to be connected to the nearest central office > 5 km away. At least one fiber must be provided for each usage unit (apartment, house, business, etc.), which is a major effort. I understand the approach with P2MP or PON and splitters to make better use of the existing fibers.
• Old POTS equipment required a lot of space and generated a lot of waste heat, so the control centers were generously sized accordingly.
• Unbundling of the last mile now also on fiber optics. ALO: Competitors like Init7 want to rent an air-conditioned, heated and dry home. This requires corresponding space in the headquarters.

I really wonder what strategy Swisscom is pursuing here. In my more rural area, entire villages and small towns over several kilometers have to be connected to a central station. I don’t know where exactly these are and which ones were even abandoned in the past due to the FTTS/FTTC/FTTB strategy.

If you were to build a new fiber optic network on a “greenfield” in a completely new geographical area, you would certainly build far fewer control centers than was the case in the “copper age”, because the distances over which an FFTH signal can still be transmitted " From a technical point of view, the data that can be transmitted “loss-free” is simply much larger than via conventional copper cables.

However, moving an existing headquarters or removing them completely will certainly involve a lot of renovation effort, so from a business perspective there will be a spread-off as to whether removing an existing headquarters will ever be worthwhile in individual cases.

But one thing is certain: from a purely network perspective, there are already more than enough existing headquarters locations for a pure FTTH network.

@millernet the ISP usually don’t need much space, a maximum of a 42 U 19″ rack. Emergency power supply is provided by the old battery storage of the EWSD in the headquarters. In the country, the headquarters are not very big, around 400m2 of floor space on the ground floor and in the basement.

Since all of the EWSD in the headquarters have already been dismantled, there is a lot of empty space available that can be rented out again with renovation work.

OLT can also be installed in small AVE ((max 20m2 area) (remote control center)).

Um, maybe a stupid question, but if I now order FTTH on demand from SC, is that also XGS-PON? Or is it AON?

@cybi arrives depending on the headquarters’ equipment. Trend more XGS-PON.

@user109 Thank you, I’ll probably find out at the IBN at the latest 😉

user109 It’s about layer 1 of the OSI model. Through direct access to the fiber to the end customer or Init7 has full flexibility regarding the usage unit. It could also “beam” quantum through the fiber 😂 or offer ⅒/25/50/100/400 Gbit/s or terabit internet, the physics of the fiber optic is basically the limit and not the layer 2 technology of it respective PON technology. With PON or GPON or XGS-PON you are bound to the limitations of the technology. The whole discussion and argument with the ComCo is not about PON vs. AON, but rather the topology (P2MP vs. P2P) of the fiber optic network between headquarters and end customers, so to speak about the discussion of where the splitters are located. Originally these were not only in the headquarters, but between feeder and drop or even in the BEP. So there is no chance for a competitor to get an ALO or Rent layer 1 access from the headquarters. You had to rely on Swisscom’s PON tree for better or worse, or you could buy your own PON tree between the headquarters and the usage unit for a lot of money as a cooperation partner (e.g. Salt). As a relatively small number, you would have had no chance of getting to Layer 1 and would have only had to purchase your services via BBCS from Swisscom and therefore via Swisscom’s technology (Layer ⅔). The nice thing about P2MP or PON is the possibility of saving fibers in the feeder. to make much better use of existing fibers. Swisscom would have to add fewer fibers and would certainly be able to expand more quickly. But that’s no longer up for debate, because a discussion about would have been couldn’t get us anywhere. The status of the old telephone network has now been restored, so to speak: one or even two direct connections (P2P) between the usage unit and the central office. PON or AON can be used flexibly by any provider in this central office. No technology is dictated.

Here is the structure at Init7:

The lower 4 switches of the access level are Cisco Catalyst with 48 SFP28 ports, where 1, 10 or 25 Bidi Gigabit Ethernet modules can be flexibly installed. The top two switches of the aggregation layer are also from Cisco and have 32 QSFP28 ports for 100 Gigabit Ethernet. Each switch is connected to an aggregation switch with 1× 100 Gigabit/s. So in this example, if we assume full occupancy with 50% 25 Gigabit/s and 50% 10 Gigabit/s (completely unrealistic, most customers are probably still on 1 Gbit/s):

• 0.5 * 48 * 25 Gbit/s = 600 Gbps
• 0.5 * 48 * 10 Gbit/s = 240 Gbit/s
• Total: 840 Gbit/s
• Uplink: 200Gbps
• Ratio on aggregation switch: 1: 4.2

Now the connection of the two aggregation switches, each with 1 × 100 Gbit/s, into the core network of Init7 comes into play:

• 4 access switches with 2 × 100 Gbit/s each (multimode AOC) on the aggregation switches: 800 Gbit/s
• Uplink per aggregation switch: 100 Gbit/s
• Ratio: 1:4

Total Ratio: 1: 16.8

Another advantage of the setup: off-the-shelf hardware and no proprietary telco equipment.

(I don’t know where these 4 DACs are going. Photos come from [https://ipng.ch/s/articles/2021/08/28/fiber7-x.html](https://ipng.ch /s/articles/2021/08/28/fiber7-x.html) )

---

@cybi wrote:

Um, maybe a stupid question, but if I now order FTTH on demand from SC, is that also XGS-PON? Or is it AON?

---

Only P2P is built. The provider of your choice decides on the technology and provides it at the headquarters. the POP or the point where your fiber “arrives”. If the control center is a POP, Init7 will use AON and Swisscom, Sunrise etc. will use XGS-PON. Many providers will offer advance services via BBCS from Swisscom and thus XGS-PON.

Whoaahh. You simply write too much, nobody wants to read who should (have to) still be working 😁

Just briefly: I cannot (and am not allowed to @user109) say exactly how our network is/will be structured.

• Much of what is written here is not entirely true.
• This means we can monitor the utilization of the PON trees through constant monitoring and use HU/KI (Human Intelligence with the help of AI) to decide when we want to expand the FAN uplinks many times over.
• The overbooking calculation here is a bit pessimistic.

About the distance:

• No, it doesn’t matter at all!
• The biggest problem is splices and the longer a fiber optic cable, the more such splices there are. Of course you try to avoid them as much as possible, but they are unavoidable if there are junctions.
• It is also known that these splices age, so the “attenuation” can increase over time.
• Further, SFPs lose their diode performance and become blind over time, which affects distance.

Strategy:

• Copper will be used for >10 years.
• In 2024 there will be a significant fiber push, even in rural areas where feasible (e.g. in my case 😁)
• There may be money from the federal government to develop areas that are not economically viable (first this has to be analyzed… and that will certainly take time)
• The development of remote areas can also be done using alternative technologies (SAT, 5G,…).
• 1 fiber per usage unit in the central office is enough (who the heck decided to install an OTO with 4 plugs)
• In the central office, this fiber can be switched to any of the ISP’s equipment. The comparison with POTS is good because xgsPON is also a type of TDM technology and only one line runs into the apartment.

It was short, right?

Many greetings and enjoy the holidays