Around 3 months ago I announced that I launched chiadownload[.]net (topic now hidden & archived for some reason).
As it turned out, DreamHost didn’t have affordable bandwidth so I switched to DigitalOcean. That was better, bandwidth was 5x cheaper than DreamHost.
As of last month, the costs of running the service have drastically increased. Perhaps this is due to an influx of new Chia farmers. The site averages around 750 downloads per month. At ~45GB, this equates to around 33TB of data transfer (billed at USD $0.01/GB).
Several months back, Chia Network’s engineering task list had several items aimed and improving blockchain sync on full nodes. One surrounding HTTP syncing can no longer be found, and Speed up initial sync is deprioritized, and Farm sooner and sync full node faster has was targeted for Q2 2022 but has not been moved to in progress. All focus seems to be elsewhere.
TLDR; I can no longer justify the costs of running the service. The site has not received any notable donations from the downloaders (it has received less than $1 USD worth of XCH today). Should I pull the plug? If no, why not? Has syncing improved for anyone in the newer versions of Chia to the point where a download site like this is no longer necessary, or is it still beneficial to the community?
Really sad to see you getting to this point. I think it highlights a couple of problems with Chia ecosystems.
The first is what you clearly outlined, about Chia deprioritizing enhancements that are obviously needed and have big support in the community but are not really high on their agenda (ATM). The obvious optics is that as long as the pain is not on Chia’s side, but rather farmers’, those tasks have really low priorities. On the other hand, if you recall, they really hurried up to close that SSL vulnerability, as that was threatening network collapse. As Chia is pushing for the IPO, it looks like some collateral damage (lost farmers) is not a big deal with the netspace we currently have.
The second issue that is rarely discussed is Chia not really supporting devs around the project. Sure, they have grant program (Chia Cultivation Grant program - Chia Network), but my understanding is that so far only Nucle.io wallet and HashGreen.net DEX got something though it, even though we are over one year into this project. (I have not heard about Max, Harold, or Gil being offered that grant (list is of course not restricted to those three, but I think all farmers know about them).) Other than that, it is mostly a graveyard of devs working on various projects. So, I would really want to know whether there is any meaning behind those words (from that grant page):
“Chia Networks is committed to growing and nurturing the Chia ecosystem in as many ways as possible, not only via applications and functionality we develop, but via encouraging our highly engaged community to unleash its creativity within the ecosystem.”
I had considered the grant program, along with ads or charging a small fee (enough to cover bandwidth costs) to download. Each have their own issues. Since only Nucle and HashGreen have received grants, it leads me to believe that program isn’t a possibility worth trying.
Torrents or an unmetered, fixed bandwidth host might be the way to go, aside from decentralizing this service as much as possible.
I was planning on releasing the source code for the download exporter so anyone can run a similar service easily and brand a website for it however they see fit but am not sure there’s much interest in that and it would have to be maintained.
I’ve never used a download service, nor would I envision I would need to as I keep a couple nodes up to date. Now this is only a guess, my opinion, but perhaps one point is that there are likely few new farmers starting with Chia. You may be facing simply that you provide a service that some few need, and of those that do, many probably don’t even know about your service (or any similar service) to take advantage of it.
As netspace has shrunk, just a guess, but likely most of the current node owners have been around a while, if not from the start of mainnet. With that, they have experience enough to know to keep an up-to-dat db on hand for any missteps that might happen.
In almost a year and half I’ve only needed to fully download the db maybe twice. Those that do need to, should have learned (the hard way) it’s not fun and so have solved their issues and moved on.
I would say your own experience needs to be your guide and to do what that tells you, however painful it may be to accept in the case you really do decide to shut it down.
Holy shit dude! You were paying through the nose for bandwidth! I have an unmetered server that I’d set you a VPS up on for ⅓ the price that you were paying! Hell, I’d be able to pay my server colo and pocket some money at that rate!
Yeah it isn’t exactly cheap when pushing over 30TB per month but it is the cheapest s3 compatible storage service I can find (way cheaper than Amazon and others). Switching to a 1Gbps unmetered server may make more sense though. The current service offers 10+Gbps transfer speeds and I don’t have to worry about the infrastructure, running a web server, installing updates, etc but that might be a necessary tradeoff.
Feel free to PM me your provider and I’ll check them out.
Moooarrr characters