Status as of Tuesday, March 27, 2018
Posted: Tue Mar 27, 2018 2:07 pm
Good afternoon!
- The current estimated date for the fiber installation is April 5. While we don't have a firm commitment to that date yet, it is more likely than not that, sometime next week, there will be a period where all services will be offline for the changeover. We'll post more about this outage when the final date is determined. We estimate that the new connection will reduce latency between Google and the server from 100ms to 15ms, along with other benefits.
- SHA-256 mining is benefiting recently due to the profitability of merge-mining Unobtanium, so make sure to join before we gain so much hashrate that the benefit goes away.
- A downside to this is that because we are mining bitcoins as the primary coin a lot of the time, the pool is almost certain to consistently lose money until a block is found. At any time, the expected amount of money we will lose before the next block is found is 69% of the block reward, or about $80,000. We still guarantee payouts within 24 hours, but the inconsistency of blocks may cause payout times to vary throughout the day until SHA-256 hashrate increases and provides regular earnings like scrypt does now.
- Chris determined that it will not be feasible to run a Ripple node in our local environment. Ripple uses a ridiculous amount of bandwidth, which simply is unaffordable. Therefore, we will rely on the public Ripple nodes that the company provides. We figure that it would be foolish for them to log private keys, and if their nodes were ever compromised, the value of the Ripple network would decline so much that anything stolen from us would be worthless anyway. He's going to work on that tonight.
- The mining servers are approaching full load. The components have arrived and Chris will double the number of mining servers around April 5.
- I've been making some changes to the mining servers over the past few days to make them handle events, like new blocks, more effectively. Currently, when a block comes in, the mining servers send out work to all miners for that block, and then move on to the next task. But when everyone is mining litecoins and a new block comes in, it takes a long time to send that work, locking up the mining servers. The new code will periodically allow other things to be done during this process, and if another block comes in before all the work is sent, then the old work will never be sent. This change will improve performance (by sending less work), allow the server to prioritize what's important, and reduce the number of work restarts sent to miners. I suspect that these lockups are also causing miners to think the server isn't doing anything and disconnect, resulting in weird network connectivity issues.