Incentives and payments for making SNOs stay

4 posts were merged into an existing topic: Why does Storj use the STORJ token?

Rather than incentives, I would opt for higher utilization on tardigrade network. As it currently is, weā€™re burning money for utility bills and hardware deterioration rather than making money.

I hope all of that will change in January when the network is fully launched.

4 Likes

I suggest to have an account that showing the actual earnings and the escrow for each month.

3 Likes

Lots of good suggestions for improving the SNO experience in this topic.
Here a small list of things that would help me keep a node up and running for long time.

  • Improving the SNO dashboard should be an imperative.
    • Historical data should be available, at least for the last month.
    • As other mention earning and escrow data.
    • Available downtime, this would help SNO manage software/hardware upgrade
  • Available downtime before getting kick out of network should be more transparent for a host.
    • Also possibility to receive email when node is down for more then x minutes would help SNO react to a failure of any type.
    • Host could (slowly or rapidly) lose their hosting data as repair occur instead of being kick-out instantly after being down for x time + pay for repair cost

I donā€™t know how this 99.3% up-time condition is implemented, if itā€™s on total time since start, last 6 month or last 30 days. I think making 99.3% uptime on the last 6 month would make sense.after 100% uptime, a host could be down for 30h in a row and not being kicked out. At the same time, if the host is running 100% on the last 6 months, my reputation shouldnā€™t be penalize for the few hours the host was offline in the first days of hosting.
I guess a 30 hours downtime even for good host, before starting the repair process could put at risk some data.
But I think 12 hours would be a minimum a host could accumulate in available downtime.

Any good link to information about update/downtime + when host is kick-out would be appreciate, I didnā€™t find that specific information.

First post, so I want to add that Storj project is amazing, keep on the good works!

5 Likes

Totally agree with @Pentium100: Home users have non-professional connections that just drop time to time. I experienced an Internet cut-off for more than half a day once, and we do get smaller interruptions regularly (of a few hours).

That would be immensly frustrating (and unfair in my opinion) to get disqualified after months and months of service, just because our ISP failed at some point. That would be very discouraging for me!
Iā€™d rather have temporary penalty (reputation, lower traficā€¦) because of a failure Iā€™m not responsible for than to be completely banned.

I do understand however that the Storj network needs to ensure that no data is lost.

What about:

  • Stop accounting storage while a node turns out to be offline
  • Quickly start to repair data (say after a couple of hours?): I imagine this process to be quite time-consuming?
  • The longer youā€™re offline, the more your reputation takes a hit
  • If you get offline for a very long time, the escrow process could kick back in progressively too, why not

If the node is offline long enough for all the data to be repaired elsewhere, putting your node back online would empty it as it would now contain only superfluous data.
And from then, its reputation should restore itself slowly with time.

But really, disqualification should be the last thing to happen when a node has been offline for a very long time, like itā€™s been abandonned (for several months).

And even then, Iā€™m not even sure itā€™s positive to have any disqualification process: If a very old node were to be put back online after months of disappearance, it could be simply treated as a new Node, with no reputation and the escrow process active.

Your views?

2 Likes

Agree @Pac and @Pentium100 about offline time.
I think 5 hours a month is too strict. My actual case:

  • 8h AM go to work.
  • 1h PM: whole electric power of my district off to maintain in 1hour.
  • 2h PM: electric on -> router reset and got a new IP (I don;t have static IP for router)
    -> Storj node canā€™t work correctly due to IP of router setting in run script is changed.
  • I see my SNO offline from work place but can not connect to change its config.
  • 6hPM: I come home and change its setting right away.
    -> I can avoid 5hours punishment but I think it is very strict for my case. If the electric outage at morning, my SNO was already removed.
    Could Storj team consider to expand its about 24h or something else?
    Thanks,
4 Likes

Apart from electricity problem the offline time seems to be your fault due to wrong setup because you didnā€™t set up a DDNS service so your changing IP address wouldnā€™t be a problem.

4 Likes

Oh, thanks, ā€œDDNS serviceā€ first time to hear.
Do you recommend any good site for this service? It is free? Thanks,

1 Like

@tldread: This depends on the hardware you use to access Internet.

Most support a certain number of Dynamic DNS services. Check which ones your ISP device supports (by accessing its admin interface), then choose one of them after making sure itā€™s free of charge.

Personnally, I use changeip.com

3 Likes

I will add that Storj said in other topics that repair traffic is expensive and that if a node with a lot of data gets disqualified, the escrow may not be enough to pay for it.
Wouldnā€™t it then be better if the node was no disqualified as quickly when going offline?

If the node loses, say, 10GB out of 2TB of data, wouldnā€™t it then be better to allow the node to upload the remaining 1.99TB of data before going out of the network? This would also allow for backups. My node fails, I can give back the data from the backup, instead of the network having to repair all of the data.
As I understand there are hashes etc used to ensure that the remaining data is not corrupted.

2 Likes

I agree with some of the ideas presented:

Dashboard

  • Summarized historical data for previous months (at least one)
  • Earnings estimation along with escrow held amount and release schedule
  • Accumulated or remaining downtime
  • Reputation view - does not need to be numeric, but for example approximated percentile of standing between all nodes, such as:
    ā€“ Top 1%
    ā€“ Top 5%
    ā€“ Top 20%
    ā€¦and so on

Downtime accumulation from previous months

  • Significantly reduced rate of accumulation, for example:
    ā€“ 30 mins per month with 0 downtime for the first 4 months (2hrs total)
    ā€“ 1 hr per month with 0 downtime for next five months (5hrs total)
    ā€“ Hard cap of 12 hours of available downtime accumulated over 9 months of 100% availability.

Alerting system

  • For any nodes unavailable for >900 seconds
  • For n amount of consecutive failed audits
3 Likes

Here is what would make me stay as an SNO:

  • Money obviously, like anyone else I presume, but let me elaborate on this:
    • Iā€™m not expecting to become rich as an SNO, but it would need to get me a minimum extra so I see it as an interesting enough incentive. For instance, if it turned out to pay my Internet connection, thatā€™d be cool.
      I think that even though the egress price seems kinda OK (20$/TB), storage price (1.5$/TB) on the other hand could be higher. I do understand that they cannot be very high because of replication, but a solid 3$/TB would be already way better IMHO.
    • There should be way to know when it is a good time to add storage to the pool of nodes: A notification on the dashboard for instance, when the disk is almost full and that the network needs more storage. No point in asking operators to add more storage if none is needed.
      What Iā€™m trying to say is that itā€™d be good to have automated guidance on what would be recommended to do depending on the current situation, for optimizing our nodes and associated revenues.
    • More generally, if I could see that the Storj service works better and better with time, and that my Node(s) get used more and more, that would keep me motivated to take good care of them. Also, if I could see that the more space we provide, the more we get paid, that would obviously keep people engaged as SNOs I think. Of course that will highly depend on customersā€¦
  • A precise and detailed dashboard that makes it clear and easy to follow our Nodes :
    • By displaying how much egress has been used during the month.
    • How much downtime happened but more importantly what penalty it caused (if any), and how to recover from it. For instance, if one of my nodes got offline for 12 hours, the dashboard could tell me that its reputation will be fully restored if it stays online for the next 15 days.
    • As mentioned by @fragamemnon it should be able to warn us when something goes wrong - however that would require the ā€œdashboardā€ to be centralized somewhere on the Internet, not only on nodes themselves.
  • As I was saying in a previous post, I wouldnā€™t want to be facing punitive situations where all my hard work could be lost in a few hours, should my Nodes go offlineā€¦ frustrating situations should be avoided, as much as possible so SNOs do not ā€œrage quitā€.
  • An online storage for my personal data, payable with Storj tokens could be nice (I believe thatā€™s already planned?)
  • Global stats are always a fun thing to see as a bonus (how many clients do my nodes serve, where they are roughly, etc.).
5 Likes

I asked about this and it seems there are plans for that. I asked specifically about the money for the storage being taken from my payment for running a node, so that I do not have to start the Ethereum wallet once a month to give back some of the tokens I got for running a node.

The satellite already knows the downtime, audit rate and my email address (part of the command line when starting a node), so it should be possible to send notifications for downtime, failed audits and old version of the node.

As for money, I understand that 1) the traffic is provided by real users and this is not mining and 2) the network is in Beta right now, so there are very few real users.

It would be interesting to see global network traffic stats though - then I could compare to the stats of my node and see whether 100kbps of traffic is normal right now or indicates a possible problem with my node.

2 Likes

Here is my story:
I became SNO starting from v2 launch, the main reason for me: help to build and use true cloud storage solution (as you know true cloud - should always be decentralized). Because all existing storage solutions are centralized and it is not a true cloud, it just a useful hosting panel with API.
Also, I would like to earn some money which can cover my expenses for my infrastructure for Storj and of course I would like to earn some extra money for my life and to growing my infrastructure.

The key answers why SNOā€™s not growing his storagenodes, they donā€™t have confidence that the money earned will at least be able to cover their infrastructure costs for the storagenode.
If SNO will earn significantly more than it spends for infrastructure it will be key reason for expand and grow SNO infrastructure. Also, this SNO will be staying with Storj for a very long time. (it something like first decentralized job).

My position, and why I on Storj way, I invest my knowledges, time, and some money for network and SNOā€™s grow because of the success of every individual of the network member will be the success of the entire network. So, it will be also my success like one of the individuals that help other individuals, and the network as a whole. Also, I love that I do.
Itā€™s a Storj way :tm:

About ideas and what can be improvement:

  • I like engineering discussions that we already have, I think it very helpful to collect opinions before implementing any features or improvements.
  • I very like that it open source project and everyone can look into commits and code.
  • I very like changelogs for releses
  • I would like to see more transparency in plans and progress (like public roadmap but with updated more frequent)
  • Storage node dashboard is not useful for advanced SNO. I prefer to use script than was written by @Krey it significantly more informative than clicking on dropdown and select satellites, then move arrow to the graph. I prefer to see a simple table with stat and graphs with historic stat.
    A lot of SNO waiting for this dashboard, but when it was released many were disappointed. It looks like was made only for finish sprint, but not for SNO. Please collect what SNOā€™s would like to see on the dashboard and redesign it.
  • Lack of documentation and recommendation for SNOā€™s especially for beginners, it produce a lot of bad experiences for new SNOā€™s and decrease network reputation. Please pay your attention into this problem. I personally also have a plan to share my experience with any SNO. I did not do it at this moment because I like a Doctor that tune your heart and my responsibility is too high to make mistakes. I just make sure for 200% that I improve a node rather than kill it.
  • When change made on official documentation no any notifications (example: beta tag for storage node). Please improve it, you can use SNO emails from storganode, simple changelog is enougth.

My proposition for thinking: We will soon see the first decentralized storage solution in production, how about another decentralized services and solutions? (DNS, Compute, CDN, VPN etc.) It will be an absolutely big change in the world.

3 Likes

While the data is decentralized, the metadata is not. If I upload a file using satellite x, I understand that if the satellite goes away, my data goes away with it too. Even though it would still be physically present on the nodes, I would have no way of accessing it.

1 Like

@Pentium100 oh wow, really? No redundancy in the satellites?

1 Like

Thatā€™d surprising, especially as Storj is planning on allowing anyone to run satellites in the futureā€¦

1 Like

@Pac I agree, the devs probably have thought about what happens if a satellite has hardware failure / data loss

So when will anyone be able to run a satellite?

1 Like

The ā€œsatelliteā€ does not have to be a single server. It can be a cluster of servers etc (and hopefully backed up), so a hardware failure should not destroy the data. However, if a disgruntled employee wipes the database or the company that is running the satellite goes out of business, the data is lost just like if it was stored on the satellite itself (which is essentially what Amazon and other do).
The advantages of this network is cost (cheaper then storing the data on the satellite itself) and possibly faster (if the nodes that have your files are closer to you than the satellite). However, IMO, it has the same resiliency as a regular storage service.

2 Likes

Just for information.
Now i working on New rich graphical and animated dashboard with support monitor walls.

4 Likes