Server re-balancing and our target server load
More than uptime we seek high reliability under fair pricing
The mission statement above is a challenging thing. Coins projects are dynamic, these change very often on the amount of required resources (amount of data sent and received, hash peaks, RAM usage, number of connecting nodes). Hence we need to keep servers with the highest safe workload to guarantee nodes with always the last valid block and eligible for rewarding.
Computers are fast, not smart! Admins hands-on are a must!
Anyway, our team oversees the workload of every server targeting: zero of Swap memory and about 2/3 of CPU cycles without a single task to process. The goal in here is to survive 100% of the expected usage peaks of masternodes to be 100% restart free (provisioning for network attacks and software fail is too expensive)! The usual VPS approach do fail on this because providers do seek maximum RAM and CPU usage as these are the most expensive machine parts, assuming that it is acceptable to drop some VPS in situation of overload or server and structure maintenance, and that lead to multiple masternodes going down.
The problem with traditional VPS is the fact that those are NOT reliable by design for 24x7 service, because they assume that if no downtime is a desired, complete redundancy MUST be used.
While complete redundancy would solve completely the problem it would double the cost. Hence, the best approach for masternodes is A SERVICE DESIGN THAT SEEKS TO STAY UP AND RUNNING EVEN DURING MAINTENANCE OR AN OVERLOAD EVENT.
Our servers can handle at least 60% more nodes than provisioned! And we set this typical low workload to guarantee your node running even during the high intensity peak events (for PoS and PoW).
Soon we will allow the user to track the super-server workload, full transparency ahead! And you guys will be able to see that we can perform expansions and maintenance without shutting a single node before having it assigned to another machine!