...snip (wall of text)
Although I quote you (@moppidoo) here, the answer below is of course for @trucobit as well.
First of all, I fully agree with your points that the current performance history feature and the insight and statistics of what you have been mining isn't as good as it could be. I've had a number of comments on this so it's on the list of feature to add.
Your insightful feedback is valuable and is interesting as it also brings up the concept from a Cloud Services point of view. May I ask what your view is on if the existing Performance History feature is improved, without the use of Cloud Services? For example that Awesome Miner wouldn't only include hashrate and temperature, but also pools, algorithm, mining software, power usage and profit? Of course a way to analyze and get a summary of this would also be required.
I do see both benefits and drawbacks with including Cloud Services in the concepts here so any additional comments on this topic is of course valuable.
Thanks for the great comments.
Hi Patrike,
I think Cloud based service extensions would be additional features with one exception: The storage of historical data. Nowadays, cloud is one of the major options to allow accessing to "data" to have a greater degree of redundancy. But I do understand there will be also a lot of infrastructure considerations, ongoing maintenance if a comprehensive set of services is to be implemented into cloud/server side processing, this may not mean the best use of your time if you were to be distracted or overloaded with maintenance issues.
So I think that it would be for a start, to have (like a lot of mobile apps) the Remote Agent to have the functionality to collect data ongoing basis, and perhaps, compress/upload them to AM cloud, or even allow user to choose whether it's Google Drive/Onedrive/Amazon/Dropbox...etc.
AM Host, on the other hand, have readily access to these uploaded data (and because it's in the cloud, configurable, meaning when users want to transfer configurations to say, changing the AM host for whatever reason, they will still have access to these data fairly easily as it is already backed up in the cloud) can download them locally, append, aggregate and do all the analysis/reporting + use for profit switching purposes / rule triggers.
Right now it's not uncommon for farm operators to have multiple geologically different sites due to whatever restrictions to space and cost each are facing. While setting up private VPN services allow AM host to seamlessly control the remote rigs, this also has no redundancy when say, either (remote or host) sites' connection are down due to outages. I personally experienced 4 major multi-hour outages, 3 on host site, 1 on remote site since beginning of this month, each time lasting for more than 8 hours. The intermittent up and down (or complete drop out during those time) means I have almost no control over what I'm mining on the remote site, even when the remote site still has internet connection.
This is where I think Extended Cloud service comes in, a way to provide 2nd endpoint for remote agents via AM Cloud. It does mean that the fundamental AM profit switching engine probably has to be integrated into the cloud platform though, much like HiveOS, Minerstat...etc, though I think AM should do it as a Hybrid Model, that there's a way to implement in a simplistic manner.
Such that, AM Host PC will always be the primary endpoint for controlling Remote Agents. But should connection fail between them, the remote agents could be pointed to the cloud on-demand or automatically via detection to set the Cloud as the controller endpoint for x interval (like how ignore pool, failover...etc works)
So the Cloud services doesn't have to be complicated, but should contain the profit switching engine that is the basis of AM Host, sync with AM Host profit switching settings, and take control over rigs when required.
Right now, it works like this:
Cloud <--> AM Host <--> Remote Agent
and I think it would be great if it could be made to like
Cloud <----> Remote Agent
^
|
v ^
AM Host <-----------|
(sorry, bad ascii art.....hope you get the picture)
***
This is going to be immediately helpful in one scenario, say one needed to be absent from the AM Host site for extended period of time (days, weeks) and whether the Host site has to be temporarily shut-down or simply due to outages, he/she can still have all the primary controls over remote sites without the need to reconfigure VPN, AM Host....pools, Profiles, most importantly, the down time...the more miners you have, a glitch or hiccup in transferring AM config file, the longer the operation may be interrupted. This, in contrast to say, Minerstat, means all rigs are still fully self-managed, and compare to HiveOS (afaik) still retain the best and most flexible profit switching controls.
***
apologies for another wall of text, the above is just one of the possibilities of extending AM's functionality/redundancy via the use of Cloud Services, I'm sure others could come up with more ideas on how this could be utilised. However, I believe, that a good plan is probably to keep focusing on developing AM Host as the main controller (features / stability / integration .... etc) and the Cloud, to be utilised further than just a web frontend from its current form, as a major backup controller. Feature and function wise, it can be implemented gradually over time as it matures / demand and resources allow. What do you think?
Best Regards,
). Everything involved in profit switching decision making is revolved around no more than 24 hours of data and can only provide estimate with limited samples, this is true not just for AM, but almost every collective mining management software out there.