This page is work in process. It's quite unfinished.
I understand that the approach of pplznet is very uncommon and controversial. People tend to have doubts about it. This is where I list the most common doubts in order to address them.
It's a common doubt that an infrastructure based on pplznet will be more expensive than a common approach. We would have to compare it to fiber solutions like FTTB or FTTH. In either case pplznet is much less work to establish since much less cabling is needed, also much less underground work.
Apart from that, one has to take a look at monthly and one time costs. Let's assume there is a town with 1.000 households renting a classic Internet line for 30 € a month.
1.000 customers x 30 € = 30.000 € / month
Let's assume these households agree switching to pplznet. This is a cost prediction, it's not based on empiric data.
| item | amount | price per item | total price |
|---|---|---|---|
| 10 G Uplink | 2 | 3.000 € | 6.000 € |
| /48 prefix | 1 | 0 € | 0 € |
| electricity VPN server 250W | 2 x 250 kWh at 35c/kWh | 63,00 € | 126,00 € |
| Leasing VPN Server | 2 | 150,00 € | 300,00 € |
| Rackspace rent | 2 | 150,00 € | 300,00 € |
| Node leasing for 36 months | 1.000 | 250.000,00 € | 250.000,00 / 36 = 11.111,11 € |
| total | 13.670,44 € |
| item | amount | price per item | total price |
|---|---|---|---|
| splicer | 1 | ? | ? |
| fiber cable tester | 1 | ? | ? |
| item | amount | price per item | total price |
|---|---|---|---|
| Fiber Twist Genxis Terminator | 2 | ? | ? |
| Fiber Cable | length differs | ? | ? |
pplznet will need some underground work if for example a street needs to be crossed. There are various ways to achieve that. The costs need to be distributed to the pplznet users.
If such a network gets big enough it will have to handle plenty of traffic. In doubt is the question if the nodes will be able to handle the load.
Mesh based networks have been in the field for a lot of years now. Most of them use radio connections and are not destined to deliver high data rates. Based on existing environments we can expect it to perform quite well but we will have to collect data from real environments.
Traffic to the internet and to other nodes in the cluster will run through the neighbours node. In doubt is the question about the privacy of this transmitted data.
This problem is not new to digital networks at all. Especially when using the internet your data runs across a whole lot of nodes and probably will very often cross different countries. To solve the privacy problem one shall use secure protocols. Just to name some insecure protocols to avoid:
Today the vast majority of protocols being used are secure. This means no matter who is able to intercept the communication data, will not be able to read or manipulate the content.
Today the most common mistake is not to enable encryption, when it is available. I see that mostly with E-mail clients. Make sure to activate encryption, which every good provider offers.
Within the so called secure protocols there are still some differences. For example an HTTPS connection is encrypted by definition. For the kind of encryption there are some different standards and depending on what standards are choosen, the connection will be more secure or less. It depends on both sides of the communication, the server and the client.
Some people use VPN to address that problem. Typically people choose to buy a plan at a VPN provider and then they think they are safe, which is fallacy. Your traffic shall be secure all the way, but a VPN will only protect it for a portion of the duration. Internet communication traverses several points between client and server. The traffic between your VPN endpoint and the the VPN Provider's endpoint in the datacentre will be encrypted. The rest of the way the VPN encryption is gone. What this approach actually does, it moves the point where your traffic will drop into the internet from the point where your ISP is to te point where your VPN Provider is.
The topic of communications security is too big to address here in its entirety. The main problem remains the same, no matter what kind of network you use.
Possible doubts:
Let's assume people generally want Internet connectivity. Based on that axiom there are plenty of possibilites how to achieve that goal. Today in most cases private companies deliver a product for us. Those products often grew historically. Products for private customers are stripped down as the funding of the providers is generally done by enterprises buying enterprise products. Those enterprise products offer things we do not get as a private customer:
The pplznet approch will devlier those attributes even to private users, without rising the financial barrier.
Especially in areas that lack fast and reliable Internet products it's the pplznet idea is much easier to grasp.
Private companies refuse to invest, especially in rural areas. They would have to invest a huge amount of money while the numbers of potentail clients is siginificantly lower than in urban areas. Sometimes the government closes this investment gap, but even in a developed country like Germany, we have one of the lowest rates of fiber connection of the whole world. We learn that we cannot rely on private companies neither can we on the government. Let's take into consideration that we could satisfy the need for Internet connectivity with a community project.
This opens up new possibilities that private companies do not have. The fastest way of creating connectivity between the citizens is to choose the shortest path, which is a fiber connection between two residents. This of course won't get you any Internet connectivity. But when it comes to that question, it's only about getting a fiber cable to the town, not more. The more relevant question is how to distribute this uplink to its inhabitants.
Today companies typically distribute this uplink to their users by building a star-like topology. This is extremely expensive due to a lot of reasons. Plenty of underground work is necessary and the amount of cabling is huge. There is so much redundancy in cabling, yet the Internet links offer no redundancy whatsoever.
List the downsides of using PPPoE. Point out why using plain Ethernet is adventagous.