search

tomeshnet / toronto-community-network

Organization, planning, and information related to Toronto's community network operation
https://tomesh.net/
21 stars 1 forks source link

Long Term Organizational and Financial Sustainability #75

Open Pedro-on-a-bike opened 3 years ago

Pedro-on-a-bike commented 3 years ago

This initial comment is collaborative and open to modification by all.

Task Summary

Determine the long term financial sustainability of the community network to ensure ongoing operations and service its users

🎟️ Re-ticketed from: # 📅 Due date: N/A 🎯 Success criteria: Determine and agree on a sustainable financial model

WIP Terms of Use -> https://hackmd.io/zV9l1OzWTqyWH50gXLb4pA?view

To Do

benhylau commented 3 years ago

I would also add:

  • Determine a means to compensate for labour for core roles that are essential to the network (including technical and organizational roles)

This doesn't imply everyone needs to be paid an hourly wage, but there should be some way of compensating time of people who are heavily involved, whose roles are essential to the project, even if they are in the form of stipends or a way to award internships.

re: legal organizational entity and relationships with neighbours we should look at how existing networks operate and relate to their network neighbours:

re: governance and economic model we should look at:

We should also schedule calls with some of their members and familiarize with challenges and lessons from their experience.

structur commented 3 years ago

re: commons based peer production

https://commonstransition.org/peer-to-peer-a-commons-manifesto/

benhylau commented 3 years ago

15 August 2020 meeting, we agreed to review materials here and add links we feel are relevant.

Pedro-on-a-bike commented 3 years ago

Updated To Do

Pedro-on-a-bike commented 3 years ago

Would like to introduce a Simple Business Plan outline for a community based WISP. This is not to push the network as a sole internet provider, but just to look at what the potential costs would be for a community WISP in Toronto.

Starting with a super node, looking at our current equipment selection, the LAP Sector is capable of handling 250-300Mbps with approximately 15-20 simultaneous users on a 40 MHz channel. This on average provides each user with an internet connection of approximately 15-20Mbps, which would allow for multimedia streaming of two or three devices simultaneously. Industry standard for WISPs is to over subscribe 3:1 so theoretically we could have 45-60 users per Access Point. With at least one Access Point at each corner of a building, (4) Access Points would require at minimum (1) 1Gbps backhaul to bridge traffic out, either to an Internet Exit or to another super node. A 1Gbps backhaul would allow for 67 simultaneous users at 15Mbps or 200 users at 3:1 oversubscription model.

So in summary, a Super Node, with (4) LAP Sector Access Points and (1) 1Gbps backhaul should allow for 200 users to be serviced in the local area. The direct monthly cost for a 1Gbps backhaul is approximately $700, with the monthly roof real-estate lease and electricity costs varying building to building. It would be ideal if the roof top space was donating as the network access would be a service provided to the surrounding community. It would not be unreasonable to expect a minimum of $900-$1000 monthly overhead cost to maintain a Super Node, excluding labour.

Edit It should be noted that new tests have been done and a follow-up post below regarding the capability of one access point and the Super Node as a whole. In summary, one LAP Sector is capable of 15-20 simultaneous users, however to achieve a reasonable internet usage experience, 13x LAP Sector Access Points (on a 20 MHz channel) will be required to handle an aggregate of 220 users at 3:1 oversubscription ratio, with each user at 25Mbps (instead of the previous 15Mbps). A LAP Sector will not be able to handle 45-60 users simultaneously on a 40MHz channel and it is recommended to upgrade to the more expensive Rocket AP for that application.

Since internet service would be a primary service provided, another monthly cost would be bandwidth to an internet exchange to provide all Super Nodes with Internet access, colocation lease for servers, electricity and maintenance. I don't know exact figures for these monthly costs, but I would imagine for it to be in the $1500-$2000 range, excluding labour.

With the plan of setting up (4) super nodes and (1) exit, a simple summary of the network would be:

(16) LAP Access Points potentially servicing approximately 720-960 users or (4) 1Gbps backhauls servicing 800-1000 users, resulting in an approximate monthly overhead for (direct) cost of $6000 for (4) super nodes and an exit. This (direct) overhead cost does not include (indirect) overhead costs, including communications/marketing and legal, insurance, software licensing, office and storage, equipment, travel/vehicle expense, training, etc. This can easily increase the budget by an additional $2000-$4000 per month. However the biggest indirect cost will be labour. We currently operate as a volunteer group, however maintaining a infrastructure like this, full time, will require at minimum (and most likely more) 3 people. 1 Person to handle all indirect materials, such as customer service, accounting, compliance, marketing, planning, etc; another person handling all direct materials, such as installs, equipment maintenance, engineering, service calls, outages, software programming, research and development, etc; and finally another person to act as a floating support staff between these two heads. In reality, we would probably need at least 2 support staff, one for each head. Though these roles would most be handled by multiple people working part-time, it is easier to consider this as 3 full time positions. With a market salary of $80,000 each, this would increase the monthly budget by $20,000, bringing the total monthly expense to approximately $30,000.

With approximately 1000 potential users, to cover the estimated $30,000 monthly cost, each user would be required to pay around $30 per month. It should be noted that the above cost estimates are extremely high level and overly simplified. It is most likely that the costs to operate a small community based WISP would be much higher, requiring more full-time staff, volunteer time and resources, more users to spread the costs over and many other considerations. It should also be noted that the above cost analysis doesn't allocate any funds for future expansions and access point growth. It is possible to double or triple the amount of access points by adding new radio antennas, assuming there is unsaturated 5Ghz spectrum available, though we would also need to increase the (direct) costs for backhauls and exit bandwidth. Ultimately, this would allow for more connectivity throughout the various communities and increasing the total amount of users.

In any case, it would be good to get more refined numbers for the costs associated to running a WISP as these costs could be translated over and used as a tool to help plan the the operation of a community based network.

Pedro-on-a-bike commented 3 years ago

Estimation of the maximum number of clients on a Super Node, implications for citywide coverage

There are 4 main radio bands available in the 5Ghz spectrum, Unii 1, 2, 2e, 3. Of the 4 main bands, there are a number of channels available, each with different allowable power and EIRP limit specifications. It is recommended to utilize Unii1 and 3 for PtP links as they can be pushed to a max EIRP of 4w (Unii1 requires a ‘free’ licence for outdoor use). It is recommended to utilize Unii 2 and 2e for PtMP links as these bands are limited to an EIRP of 1w. It should be noted however that Unii 2, 2e are subject to DFS and have the potential of service interruptions due to radar events.

Within the Unii 2 and 2e bands, there are a total of 13, 20MHz channels available. The max spectrum efficiency of 802.11ac is approximately 7.8bps/Hz, resulting in a theoretical max bandwidth of 156Mbps for a 20MHz channel with two spatial streams. It should be noted that due to the unlicenced nature of the 5GHz spectrum, it can be expected that these channels will not all be free of noise and there is also frame/protocol overhead that will reduce the total available bandwidth of a channel. There are equipment enhancements that can improve the efficient use of these channels, such as GPS timing, so it is not unreasonable to expect a real world bandwidth throughput of 140 Mbps per 20MHz channel. With careful planning it may be possible to have 360 degrees of surrounding coverage of a Super Node with all 13 channels used to some capacity. Based on our testing at a 2.5km radius, a CPE client is capable of a TX/RX rate of 25Mbps, which is sufficient for the average internet user. With 13, 20MHz channels available, the total aggregate available bandwidth is approximately 1820Mbps, limiting 73 simultaneous clients at 25Mbps. At a 3:1 client/subscription ratio, this brings the total maximum users per Super Node to approximately 220. It should be noted that 20Mhz channel was selected as the LAP-120 is only capable of handling 15-20 clients simultaneously and with a 3:1 user ratio, this allows for each Access Point to provide approximately on average 25Mbps / client.

With a 2.5km radius, a Super Node should be able to provide coverage to a 20km square area. The City of Toronto has a square area of 630kms, which would require 32 Super Nodes to cover. Of course this is a simplification as terrain, landscape, availability of tall buildings and LoS significantly limits the ideal location for a Super Node. None the less, this is an exercise of the potential reach and connectivity of a Super Node deployment in the 5GHz spectrum. Therefore at 220 clients per Super Node, a citywide deployment could potentially service 7040 users. It is possible to improve on this number and increase connectivity as there are new technologies that can increase the max spectrum efficiency and improve on the total aggregate of available bandwidth.

In terms of operating cost, based on the 1820Mbps capacity of a Super Node, this will require a minimum of (2x) 1000Mbps backhaul connections to a datacenter. Along with rooftop leases, electricity and other maintenance costs, it is not unreasonable to expect each Super Node to have a direct monthly operating cost of $2,000. With a projected 32 Super Nodes, this results in a monthly operating cost of $64,000. As previously mentioned, a colocation service at a datacenter will be required to handle the aggregate bandwidth of all 32 Super Nodes, which is approximately 60,000 Mbps or 60 Gbps. A projected monthly estimate for this bandwidth is $2,500, plus a colocation lease of $1000 for a ½ Rack totaling $3,500 / month. Rounding up, this brings the total monthly estimate to $70,000 to service 7,000 clients in terms of direct bandwidth costs. As previously mentioned before, there are many other considerations, such as labour, compliance, maintenance, etc, so it would not be unreasonable to expect the total operating cost to be much higher than this depending on how the organization is structured and laid out. I would imagine an organization to handle a network of this size would be approximately 10-15 people.

Pedro-on-a-bike commented 3 years ago

Potential option for a long term roadmap

The long term vision and drive of this organization has always been to bridge the gap in the digital divide, improve the overall digital literacy of the community and make this easily accessible to all. A potential long term plan and roadmap for all members then is to determine a way to develop an organization or structure that will be able to freely operate and allows all of its members to be able to connect together and gain access to this community driven content or service in a sustainable way.

The primary efforts to physically connect these members and communities together is to build out a physical infrastructure by setting up wireless antennas and receivers around the city that members can connect to, known as Super Nodes. These Super Nodes can then be interlinked either by PtP (Point to Point) wireless connections if there are in range or through backhaul tunnels over the internet. The initial costs and estimates of operating and maintaining this is described above. In the end though, the primary focus is to create an network and infrastructure across the city that is accessible to all and to be able to facilitate the creation and hosting of content and services on this network with the end goal of improving overall digital literacy and digital freedom.

In terms of sustainability, one potential option is either to have a membership fee or accept donations to help fund the operations and maintenance costs for the network/infrastructure/organization that members can pay into to ensure it can sustain itself and members are able to gain access to services and content that is hosted on the network. The actual organizational structure is still yet to be determined as it can range from being a non-profit, to a coop, to some unique multi-stakeholder arrangement. However one potential option to think about of this community is an organization that is developing a means and technology to easily network and connect all of its members together so they can gain access to content, services and other community needs with the end goal of improving improving digital literacy and close the digital divide that currently exists. I am also open to other ideas as to how we can think about this community/organization and build a road map together.