-
Notifications
You must be signed in to change notification settings - Fork 6
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
For files over 100MB, they cannot be pushed to GitHub with regular git #20
Comments
The call-block-logs crawl is pretty huge. And potentially, we want to make it available to projects without the necessity to run a node. Actually, it remains a question if we truly want that as an alternative is the user running the nodes themselves. But then still, just to keep the threshold low: What are the costs of storing roughly 1TB in the cloud or IPFS. What would be the costs of bandwidth? And do these costs motivate us to either store or not store that data? Here's the current call-block-logs crawl starting from block: 11565020 -rw-r--r-- 1 root root 73G Aug 7 21:07 11565020-12000000 It's 621GB for now, with probably another 100GB (721GB) being added until block 15301264 (current tip) is included. AWS S3Storage pricing: $0.023 per GB per month Then for data pricing, we'd need to understand how many users are indeed interested in downloading this data. But the model is 721GB per user, so it's 721GB*x, where Data Transfer pricing: $0.09 per GB All information is based on the official AWS pricing page [1]. Self-hosted ErigonErigon suggests running a node with more than 3TB of SSD storage [1]. Hetzner's AX101, which we use to run a node using rugpullindex.com, totals 111.86€ monthly [3]. Their dynamic block storage in the cloud for 2TB is 97.48€ per month. Running an Erigon node doesn't require a bit of effort operationally. It's an initial effort to monitor the chain synching, but then it runs mostly smoothly without any necessary user intervention. Using neume, a node's interfaces can be blocked for the internet - with mostly neume handling the crawling (and later potentially the API exposure). Unchained Index With IPFSTrueblocks features a "unchained index" hosted on IPFS. It allows a user to directly access Ethereum data through carefully specified identifiers. For E.g. an account's entire historical balance can be accessed by crafting the identifier and then selectively downloading partial data from IPFS. This is great as the index's identifiers have a permanence guarantee [1]. It's good because anyone can host those indexes (e.g., by pinning). But it's truly exceptional as it allows a user to traverse and directly access the data through this neat identification scheme. So assuming we'd create and host such an index on the IPFS network, what would the costs be? Truly there are two options: One being to use e.g. Hetzner's Block Storage or an IPFS cloud provider like Infura InfuraInfura doesn't allow storing beyond 200 GB of data. But at "Unlimited Storage $0.08/GB", they're much cheaper than AWS S3, making 721GB unchained index cost a total of 57.68€. At $0.12GB transfer costs, and a model centered around Pinata (IPFS)Bandwidth; 7.5 TB per month. EternumBandwidth: No info. FleekBandwidth: $0.05/GB References
|
15 data packs per month would allow us to upload the data to GitHub in this repository using https://git-lfs.github.com/ until we found a way to effectively store it on e.g. IPFS. |
It looks like Filebase.com is approx. $6/TB/month for storage and bandwidth, with the first terabyte of each included in the $5.99/month sub. I think it's IPFS backed by Sia.tech https://docs.filebase.com/billing-and-pricing/pricing-model
|
Have you seen https://estuary.tech / https://docs.estuary.tech?
|
The text was updated successfully, but these errors were encountered: