Siacoin (SC) News
I would like to be able to hold my sia on coinbase as well as my UI, and any exchange. Do you think / see this becoming a reality?
You can’t argue the fact that coinbase gives credibility. Is this something that is in the works???
I've been steadily working to support the new renter-host protocol in
us, and today I think it's ready for wider testing. I'm sure there are some rough edges remaining, and the performance could certainly be improved in a few places, but the core functionality is there. The new protocol is a huge step forward in terms of functionality and performance, so
us will support it exclusively, refusing to talk to old (pre v1.4.0) hosts.
The biggest new feature is a
PseudoFile object that supports standard I/O operations. For Go programmers, this means you can
Write a file stored on Sia just like an
*os.File stored on your local disk. The encryption, erasure-coding, parallel transfers, etc. are all handled for you transparently. For a real-world example, just take a look at how little code it takes to implement the
user HTTP server. Now, there are some serious caveats -- namely, you can only
Write to the end of a file, and proper buffering is required for good performance -- but these will be addressed as the implementation matures. Once random-access
Write is supported, I'll be releasing fully-featured FUSE bindings. Maybe I'll run
siad on a FUSE mount as a demo. :)
Other new features:
- Thanks to the new renter-host protocol, we no longer need to store any Merkle roots locally; all modifications are proven by the host and verified by the renter. So the
.contractformat can shrink to just a host key, renter key, contract ID, and the most recent contract revision. (Technically, we don't even need the revision, since we can request it from the host at any time, but we use it frequently enough that it's worth it to keep a local copy.) It's small enough to transfer via QR code, which could be a cool way to get contracts onto your phone.
- The file format was tweaked slightly to use XChaCha20 and randomly-chosen, sector-specific nonces. This means that the encryption is actually stronger than deterministic disk-encryption ciphers like XTS. Briefly, if you make a change to a file and then revert it later, the host can detect that you reverted the change if you use XTS. Using random nonces prevents this: the host can't distinguish between you reverting a change and applying an entirely different change.
- I implemented a barebones host (called
ghost) for testing purposes. It's useful if you want to quickly spin up a bunch of ephemeral hosts. It's not very mature yet (all data kept in memory, contracts aren't validated, doesn't charge anything for storage), but I would like to flesh it out more, especially if there's interest from other devs. Maybe it could become something like an antfarm for
If you're interested in using
us for a Sia-related project, reach out to me on reddit/Discord/email and I'll offer personal assistance. I'm especially interested in what the pain points are in Sia development and how we might address those. If something is difficult or unintuitive, I want to hear about it!
|24h Low / 24h High||$0.00301 / $0.00319|
All time high
|All time high||$0.09287|
|All time high date||2018-01-06 (one year ago)|
|Since all time high||-96.69 %|
|24 hours||-1.94 %|
|7 days||-1.42 %|
|14 days||-5.99 %|
|30 days||11.63 %|
|60 days||30.46 %|
|1 year||-83.95 %|
|Merged pull requests||281|