Upload speeds are a problem

People,

LIke others, I have been hanging out waiting to get onto the live SAFE Network and I am interested in doing a lot of farming and developing a number of apps - but there is a problem - the current Australian Federal Government and upload speeds. If the current government hadn’t scrapped the previous government’s Fibre to the Premises (FTTP) plan, I might not be complaining about upload speeds today. However the situation is as it is and I have to find a way of doing what I want in the current environment. For a summary for Australia (but I think it is relevant generally?) see:

At home I currently have ADSL2+ and that gives me about 14Mbs Download and about 1Mbs Upload but if I want to do a LOT of farming, my upload speed is going to be a problem. I have been looking at solutions including bonding other services like NBN wireless or 4G to the ADSL2+ plus but none of them are particularly attractive - however I will see how things go once we are live.

Anyway, this brings me to my point - isn’t it the case that MOST people on the SAFE Network are going to have much faster download speeds than upload speeds? Isn’t this a problem if SAFE wants to be “Internet 2”?

Thanks,

Phil.

3 Likes

I think that is going to be fine.

Instead of downloading one file from a massive server someplace you will be downloading small chunks from many computers spread all over the world. Plus you have caching…

Farmers with slow upload speeds are probably going to not compete terribly well against farmers with big upload pipes, but other factors will also be in play… Latency is going to make your machine closer to some requestors than other nodes…

@jreighley,

OK, I guess we will start to see soon! If 1Mbs is not sufficient I can consider the bonding options but of course that adds expense to the setup. Another way around it is to bond multiple ADSL2+ lines together but I don’t want to do that until I am in new premises and can justify the cost for new projects.

I guess it will depend on the value of SAFE coins as to whether it is worth the cost of setting up Digital Ocean servers for dedicated farming . .

Thanks,
P.

A lot of people in my country are getting fiber connections. I myself have 250Mbs up AND down for 55 Euros a month. For another 10 a month I can upgrade to 500/500 which I will probably do when the network is live. The only question then is what does my ISP consider a “fair use” amount of data. I guess for me this won’t be a problem since I only have like 2TB to share and I won’t be sending the entire 2TB around during a month. But if you want to be a huge farmer this might become the next restriction. Assuming I serve the network at full speed 24/7 I would be able to almost send the 2TB twice in 24 hours.

I feel for you guys over there in places where the upload is only a fraction of the download speed. A few years ago it was the same here, but the fiber changed that.

1 Like

@Krekc,

Damn! That’s a pretty sweet deal you have there!

P.

I HOPE there are limits to how profitable it is to do huge farming rigs with a bizillion Vaults. I’m all for setting up a number of extra Vaults in order to take advantage of the more favorable availability of safecoins early on. But the idea is that lots and lots of people provide resource, and lots and lots of people earn safecoins in lots and lots of areas all over the world, using resources that they already will be having, which cuts down on overhead.

ButI guess it’s fine, really, either way. If lots of people start out pumping in lots of resources, that’s good for the security of the network at the start. Long term profitability doing that is, I’m thinking, another matter. There are a lot of interesting inputs that start to boggle the mind.

What if data centers decide to get into the game, since they have lots of spare resources sitting around? If it’s even close to profitable to rent Digital Ocean droplets (or the like), think from the perspective of the marginal costs to large data centers. But, then again, they WOULD BE providing lots of resources to the network, so should get the benefits, just like us wee folk who can only do . . . less. If such interests do turn a significant portion of their resources to the farming, it will make for a very strong network early on, and quite profitable for them at first. But over time the democratizing effect of the network’s design will start to erode their model over time.

We’ll all do what is in our power and economic sense to do, and a true market will balance it. Don’t think there’s anyway to have a market within this design that’s not a true market.

I’ve probably strayed pretty far from the topic, but seemed worth saying. If anyone wants to chase the rabbit further afield on this, we can fork the topic.

2 Likes

You kids and your fiber optics. I remember the days I was dealing with dialup. I APPRECIATE 1MB/s

5 Likes

Does running more vaults make better use of available bandwidth? As in if I share 1 vault with 2TB it would only handle 1 request at the same time where if I ran 4 vaults with 500gb each it would be able to serve 4 requests at the same time?

@philip_rhoades & everybody else who might be interested.

[Maybe it’s worth it to try Scaleway.com][1].

€2.99 /month for 2GB Memory, 50GB SSD Disk, 1 Reserved public IPv4, 200Mbit/s Unmetered bandwidth. Maybe it would be helpfull, if a community member could put up an SAFE Network image (not me I’m not a super computer wizard)
[1]: https://www.scaleway.com/pricing/

5 Likes

@19eddyjohn75,

Very interesting!

Thanks!

Phil.

2 Likes

I’m sold. Definitely need a tutorial, though, on how to install vaults.

One could easily run 4 or more vaults on this type of spec, seems to me.

1 Like

Here is something I’ve been thinking about for awhile. I would like to get everyone’s input.


There may be a huge market for a “Rent-A-Vault” business…

Demand
There are people who want to “professionally” farm Safecoin but lack the resources/ability/location to do it efficiently. The OP is an example shared by many in this community. This is not meant for casual farmers.

Obstacles

  1. Cloud servers already have infrastructure in place… may provide lower cost.
  2. It should be easier to buy Safecoin on an exchange.
  3. Centralization of farming.
  4. Heavy responsibility for maintenance… making sure the vaults are online 24/7.

Benefits

  1. Renting a vault means your wallet address goes in the vault, and verifiable online.
  2. No need to install anything, just click to rent a vault and start receiving Safecoin.
  3. Vault statistics shown, allowing potential renters to see which vaults are performing well… attracting a higher premium.

Pitch
High bandwith ISP service cost around $50-$100 per month. If you’re professionally farming, you would likely want a dedicated line, so your own browsing doesn’t use bandwith needed by your vaults… remember they also have to cache data for free. More vaults means more caching and consensus processing.

Catch
I have some community members in mind to form a collaborative union to combine resources (Fiber Locations, Fiat Money, Hardware, And Maintenance Managers)… In other words, a professional farming club, where we find and use the best of the best to get the maximum benefit out of farming.

If I can put it together, would you be interested in joining?

5 Likes

Seems like it would be much easier to rent space on existing infrastructure… For example, I already have a dedicated virtual server at my web host, nowhere near capacity…

Obstacle #1 and #4 make this pretty much a slam dunk… If the vault is profitable I can increase the size or quantity on demand with very little work.

You eventually could probably even DAO the whole thing… And have the vault rent themselves. Rent more when it is profitable to buy, rent less when it is profitable to sell… This would be more effective than involving those fickle humans…

Most cloud servers (Rackspace, Amazon etc) have an “install wordpress” or “Install Joomla” etc. Install SAFEvault would be just as easy, and most likely could be set up as a turnkey within an hour or two for the whole world to rent on demand.

2 Likes

Agreed and acknowledged.

I like your idea of DAO for autonomous vault rental using cloud servers.

I’m not against any solution, as long as it works.

The greater issue would be that if AWS (for example) is the most economical farming model, then it greatly centralizes things – and if AWS goes down, MaidSAFE may have trouble – particularly if there are a significant number of chunks where 4/4 copies are stored on servers dependent on such a service…

I don’t know that the network could really see or identify such a risk.

It may be useful to reward “Uncles” like etherium does - so that diversity is maintained and you have slow (and most likely decentralized) hardware backing up the fast (and most likely centralized) hardware…

1 Like

@dyamanaka,

I am most definitely interested. My goal for the next five years is to get my apps and farms going on SAFE and to move as much of my digital life there as possible. We might be able to avoid the centralisation to some extent by choosing (automatically?) the locations where vaults are created - all the cloud providers have a number of physical locations for their server farms.

I already have a Docker image so that part is easy too - spinning vaults up and down on-demand should be no problem?

My main project that I need to get going to recover some cash involves building a 12-house village:

http://lev.com.au

My intention, in the first instance, is to rent the houses as they are built (for possible sale later) but if I can get NBN fibre to the premises (or at least multiple, bonded ADSL2+, wireless / WiFi connections), then part of the deal would be supplying an Internet service in each house - where the device in each house would, of course, also be hosting vaults.

P.

2 Likes

This all assumes that the number of coins farmed actually reaches the cost per month. Maybe early on it will only be for people who wish to collect coins for later on.

Early on there may occur (or not) the situation where people are PUTting a lot, but GETting little, so farming rewards are small. Those wishing for dollars would disappointed if this happened, those wishing for coins to collect/use in PUTting maybe happy.

For the home farmer, it is likely that the monthly cost is zero, or very very small, so early on the rented space will only compete if bandwidth, speed, ON time are an issue for (extra) home farming.

Later on when the network has matured a little and the value of safecoin is a little higher then the cost benefits will increase and the rented vault may even become viable.

The biggest question with this is, that like the “super smart investor” who asks people to join him in making a killing, is WHY. If the “super smart investor” is SO good then why ask for money from others to dilute his eventual gains. He would do it alone.

So like this, if farming is so profitable and you have vault renting business then you need to charge more than your costs and if it was so profitable to farm this way that others could make money after your added costs then WHY invite others??? It would be more profitable for your group to farm for themselves and keep 100% of the rewards (less the costs). Otherwise you would be leveraging on others to receive more from them than you would get from farming.

In other words why would anyone trust that they will receive more than they pay you. Perception is that maybe some lucky ones might but on average they won’t. If all made a profit then you would just use the vaults for yourself and make more money/coins.

Better and more trusted if you posted guides of the better places to rent servers and how to install vaults. Maybe even one off charge to do the install. That way the (middle-)middle man is removed and maximises profits for the end user.

I detect a significant amount of cynicism.

Some people need help because they lack certain elements to farm at max potential. I proposed a way to collaborate resources to address those weak links… partnering those with Fiber access with those who have money to pay bills, or hardware to contribute.

If people don’t trust each other, they can go at it alone. For everyone else, the invitation is open. I’ll be opening a private thread to discuss things in more detail. Thanks

1 Like

Actually it isn’t as it sounds. I do like the idea, but this was a Con and because people adopting SAFE will include more than its fair share of people who do not trust, I thought it should be mentioned.

So you are saying that it is from the goodness of your heart that your group would do this? that question sounds like criticism but more an honest question for yourself to answer, because I’ve seen this before and human nature ends up defeating those who do not answer that question honestly.

If it is from the right motives then it will work.

I also see that offering “hand holding” services as another way to help people. They don’t have to trust you not to change the pay-address for 3 days a month, they feel more in control and can add more if they wish.


On another note, have you considered the effects of all the vaults being seen by the network as living within a local net, and in your case living on one IPv4 subnet . There was discussions that having a lot of vaults on a local net would end up reducing the relative number of PUTs that occur into each vault.

Not sure of what the devs eventually said or did in that respect, but if local nets do reduce the rate of data being stored on each vault then concentrated servers farming might end up being counter productive the more that are added.

Perhaps @neo is being cynical, but he’s got some good points, regardless. It really is hard to figure the math on it at this point.

I’m attracted to the cloud solution to a certain degree because I’m on the go a lot, and having a dedicated could setup that I can monitor from anywhere may be more profitable than the home service which I can’t always monitor and put back up if there’s a problem.

Put there is the matter of low marginal cost if people are using resources they don’t have to pay up much extra for. On the flip side, I’m not unwilling to throw $60 at it for a year of running a couple or three vaults on a cloud droplet, just to see what happens.

But the real profit in the network is going to be made by app devs and content providers, I think.

3 Likes