Making Bank
-
I can't say I'm a fan of things like your host crashing being nearly enough to lose revenue.
I understand that you target mainly enterprises, but personally I wish this could be successfully used by single individuals willing to do tit-for-tat file storage... so instead of having strict uptime requirements, couldn't you move towards some system where uptime gets measured and rewarded, and/or renters can decide whether to host with high redundancy on hosts with low uptime, or with lower redundancy on hosts with high uptimes (or automatically based on the ratio of these)?
-
Unfortunately the redundancy requirements get very extreme one uptime drops below about 95%. With 95%, we can get by with like 2x redundancy. With 98% uptime, we can get by with 1.5x redundancy. At 80% uptime (still a lot!), you need like 5x redunancy. At 50% uptime (which is more than every night, it's most of the day every day), you need like 25x redundancy.
The penalties for being offline every now and then are not very big. If you have 95% uptime, you'll be ~85% as profitable as someone with 100% uptime. It's not a huge penalty. But at 85% uptime you'll be like 40% as profitable, which is a huge penalty. And it's structured like that because hosts with less than 90% uptime are basically worthless to the redundancy scheme. They add very little value and you need a lot of them to be fault tolerant.
-
It's a shame, but it sounds like you've done your maths on this, so I can't really dissent...
Except, extreme redundancy might possibly be less of a problem than it seems, if storage on each "unreliable" host is a lot cheaper than storage on "reliable" ones, and the world broadly speaking has a lot of unused disk space.
I guess it comes down to my general thinking that it would be best to try and let parameters (like price, redundancy, penalties...) change naturally instead of hardcoding them.
-
A competitive price for storage, assuming you are buying the rigs yourself, is about $3 / TB / Mo. $5 / TB / Mo tops, and that's assuming you've got a bunch of overhead costs. Some setups might be able to get the cost as low as $1.5 / TB / Mo.
So, hosts with 80% uptime, or 50% uptime are going to need to compete with that, meaning they are making closer to 50 cents / TB / Mo for storing data. That's not really incentive enough for me to put my laptop on the network, especially considering I'd be offering far less than 1 TB. So... we really do rely on hosts that can have high uptime. I just don't think it makes sense to cater to people that can't be online most all the time (with some leniency for reboots and maintenance).
The one place that low-uptime hosts might be valuable is bandwidth. Depending on how bandwidth-starved the network ends up being, it might make sense for hosts to take some files for free and then have expensive prices for the downloading of those files. If they are popular files, low uptime won't be an issue because it's not a question of availability or reliability, but rather a question of throughput. But that's something we'll be looking at more down the road.
-
You're concentrating on monetary cost, in "real dollars", here, but I'm thinking also about the user who doesn't want to get into a virtual economy and buy strange coins with their credit card; who, instead, wouldn't mind giving up some of their storage space simply in exchange for other people's storage space to backup data on.
In this tit-for-tat usage, a user may not even be aware there is a currency and transactions involved (or at least, that depends on how transparent software makes it): they'd simply take part in a worldwide peer-to-peer network where storage is distributed and "load-balanced" all over (but, contrary to some existing ones, on where cheating is made impossible by the blockchain).
(By "load-balanced" I mean that if I need more storage today, and you need more storage tomorrow, we can handle that by lending and borrowing, instead of each of us buying more hard drives.)
-
How much is your highest host/farmer paid a month in dollars? I've seen our highest farmers be paid anywhere from $40-$70 per month.
-
@super3 said:
How much is your highest host/farmer paid a month in dollars? I've seen our highest farmers be paid anywhere from $40-$70 per month.
How are the payouts calculated? Since Storj right now doesn't have an actual market for people to trade storage space, what does this number mean?
-
How much is your highest host/farmer paid a month in dollars? I've seen our highest farmers be paid anywhere from $40-$70 per month.
I'm assuming that money is coming out of Storj's pocket? $3,000 / mo is not too bad of a burn rate to keep that many hosts online, but it's also worth pointing out that a lot of those payments were calculated after a huge jump in the value of SJCX. At the time of the payout, was the value that high? Was the promise that high?
The most expensive hosts on the Sia network are able to sell their storage for about $4 / TB / Mo right now, but none of them are serving a full TB of data. The whole Sia network currently has about 1.5 TB of active data.
-
@TacoByte said:
@super3 said:
How much is your highest host/farmer paid a month in dollars? I've seen our highest farmers be paid anywhere from $40-$70 per month.
How are the payouts calculated? Since Storj right now doesn't have an actual market for people to trade storage space, what does this number mean?
USD value of SJCX paid at time of payment.
@Taek said:
How much is your highest host/farmer paid a month in dollars? I've seen our highest farmers be paid anywhere from $40-$70 per month.
I'm assuming that money is coming out of Storj's pocket? $3,000 / mo is not too bad of a burn rate to keep that many hosts online, but it's also worth pointing out that a lot of those payments were calculated after a huge jump in the value of SJCX. At the time of the payout, was the value that high? Was the promise that high?
The most expensive hosts on the Sia network are able to sell their storage for about $4 / TB / Mo right now, but none of them are serving a full TB of data. The whole Sia network currently has about 1.5 TB of active data.
This comes from the pool of SJCX that are unreleased. The main value of SJCX is to reward farmers. We don't have any mining, so it works out quite nice and non-dilutive.
We release 100k SJCX each month regardless of value. Not really. We have been doing this since September. The value has gone up but so have the number of farmers.
-
Where does the 100k SCJX come from?
Also, what's the ownership/distribution model for StorjCoin? It is premined?
-
@super3 said:
How much is your highest host/farmer paid a month in dollars? I've seen our highest farmers be paid anywhere from $40-$70 per month.
One hand, Storj is burning their own funds to incentivize farmers. On other other hand, the Sia network storage is underutilized. So, the two are clearly not comparable at present.
If a Sia host wants to aim for 100% utilization, it needs to be price competitive. Amazon'z Glacier pricing is about $7 TB/mo. If a host wants to be competitive with that, he needs to take into account redundancy (lets jump ahead and assume the network has grown to 100+ hosts and redundancy has dropped to x1.5). The host then needs to charge no more than $4.67 to remain competitive. We assume the host automatically adjusts its price in Siacoins automatically according to prevailing market conditions (and that over time the exchange rate is stable). We also assume 100% utilization and no lost contracts.
A host with 25TB storage then stands to make $116 USD/mo. That equals about a hardware investment of about 5 TB. So, the host could continue growing its storage at 5 TB/mo while remaining at break-even (ignoring the initial investment in hardware).
So, it stands to reason that the network's capacity would continue to grow as long as the storage is utilized. And, if the protocol is scaleable, the network fast and price as good as the alternatives, I see no reason it would not.
-
Amazon'z Glacier pricing is about $7 TB/mo.
Sia does not compete with Glacier, Sia competes with Amazon S3. So the competition point is actually >$25/Mo. Meaning hosts can charge as much as $15 / TB / Mo (at 1.5 redundancy) and still be competitive (in the long run).
So, it stands to reason that the network's capacity would continue to grow as long as the storage is utilized.
Yes, I believe the number one reason we aren't seeing more hosts is that there's no utilization. And the number one reason we aren't seeing utilization is because the software isn't ready yet to handle massive volumes of files. But it will be soon :)
