Witness Update - RPC Node + Blocksize parameter update

in #witness-update7 years ago (edited)


We want to thank everyone that has supported @curie's witness so far. Curie is currently witness number 6. We appreciate all your votes and are determined to serve the community in every right way possible. Please remember that @curie is a community-driven witness and all rewards gotten from our witness are returned back to the community one way or the other.

For all witness related queries, consider joining our discord server and visiting the #witness-enquiries channel where your queries will be answered adequately.

  • For those new to Curie, please follow @curie, and join us on Discord: https://discord.gg/jQtWbfj

  • Follow @curie's votes to support the authors. Please consider following our trail and voting for curated authors. If you are a SteemAuto user, @curie is an available trail to follow.


RPC Node

At Curie our focus is and always will be community development, support and the sustainability of the Steem platform. In accordance with our core focus, we have brought up a full node for further node reliability to support the Steem platform.

We’re running our full node on 256GB Ram which is enough to hold most of the shared_memory in-memory, hence making our full node fast and reliable.

Our fullnode is available for API requests at: https://rpc.curiesteem.com/


Specs:

20x Core Xeon architecture CPU, 256GB Ram, 1.5TB SSD, 1Gbit Dedicated Port.

In the coming weeks we plan on bringing up a backup full node and implement load-balancing such that we can sustain 2 of these full nodes, node balanced to our node RPC point in order to make the API point significantly more reliable for future use as the RPC point becomes used more frequently.


Blocksize parameter

It has become evident over the past month or two, that we are testing significant bandwidth limitations due to the current blocksize of 64kb, due to this we ran a vote amongst our curation team, reviewers and operators in order to decide whether it would be best to stick with a 64kb blocksize, or to move to something slightly larger (like 96kb or 112kb) which would allow for newer users to not be restricted by bandwidth in the future.

It was unanimously decided that blocksize should increase with all information provided, however, the blocksize most voted for was 96kb and hence we have updated our witness blocksize to reflect this consensus.

We suggest any current top-witnesses (who many have discussed the situation either in posts/elsewhere) to take a look at blocksize and bandwidth to make an informed decision with respect to the future growth of the Steem platform about their current blocksize parameter.


For more information about @curie and all its activities you can read our whitepaper https://steemit.com/curation/@curie/curie-whitepaper or check out the pdf file at http://curiesteem.com/#whitepaper, vote for @curie as a witness and follow us @curie!

Sort:  

Afaik, 256 GB RAM for a full node is not enough anymore.

It is with SSD based Swap, I think GTG still has his full on a 64gb/128gb server with the rest on SSD Swap IIRC and is still one of the most reliable RPC nodes (In my experience). In fact the blockchain is only just approaching 300GB in terms of shared_memory on the RPC nodes, hence a 256GB RAM server with ~400GB Swap Should suffice API requirements at relatively high speeds for quite a while, if it is noticeably a problem with respect to API call speeds, we will opt to either upgrade to a 756GB+ server, or run multiple servers for different APIs and use Jussi.

How can I support the witness.... Just got this now and I wanna render little help to

Witness Voting is at:

https://steemit.com/~witnesses

You can give curie a vote there.

Number 6 is great. This is goodnews. We can push this through to number one.

Oh, wow: I did not know you were able to adjust block-size independently on each node like that... Is that what you're saying? One node may be mining blocks that are 64kb, while others (yours) mines 96kb-sized blocks?? Please correct me if I'm wrong.

Additionally, thank you for your post--it is always nice to feel informed / kept abreast of the situation--especially one which holds such import as a Steem Witness!!

My understanding (and @locikll please correct me if I am wrong) is that it is consensus based - all witnesses set these parameters more or less independently and when a critical number of witnesses has raised the blocksize parameter than it is implemented blockchain wide.

Interesting... so.. each node sets its own parameters, but if said parameters reaches some certain aggregate threshold, then all nodes adjust to this (new) "threshold?"

that is my understanding, that once the consensus threshold is reached the new parameter would be required and witnesses not supporting it would not be producing blocks.

Great, @curie. We hope your later position will be better.

  1. it's a good thing.Definitely minimize the problem of bandwidth if you download now you will move to the beginning of the list.What happens to the water problem, solve the problem

6th witness, wow kudos on that... Very soon curie will be top 3 on witness list, i know it will happen 󾠮󾠯󾠰

I think number 6 is a great achievement. However i am convinced that sticking to your focus of community development, especially with this awesome introduction of your full node for the stability and welfare of the platform, i think you'll hit number 1 sooner than we think. Success @curie