Over the last few months, we’ve made significant changes to our infrastructure based on lessons learned over the past 2 years of testing and optimizing hardware for EOSIO purposes. Our hardware is optimized for the specific requirements of the variety of services we offer and we want to take the time to share what we’ve been doing and the positive impacts it will bring to many of our services.
EOS Nation makes infrastructure investment a priority, which is reflected in the highly performant hardware we use to continuously offer world class services on the EOSIO networks we serve. Our bare-metal hardware spans multiple data centres and we use different network providers located in multiple jurisdictions for maximum redundancy.
Our latest changes come in 3 parts: data centre space, server infrastructure, and network bandwidth. Let’s have a look at each in more detail:
Data center Space
The data centre where we hosted most of our services had limited space and was not cost efficient, but was otherwise working well. We now moved our primary data centre to a new location with much more space for us to grow, while reducing our costs at the same time. We now use twice as much space and have room to grow some more as needed.
Our block producing nodes have been using Intel Core i9-9900K processors for a long time. We’ve now upgraded to the newer i9-9900KS for all of our block producer nodes across all of the EOSIO networks on which we produce blocks. This CPU change increased our clock speed from 4.6GHz to 5.0GHz. And while we were at it, we also doubled the RAM on the i9s to 128GB per server.
As explained in a previous article, we run private P2P “blocks” and “transaction” nodes to exchange blocks with all other block producers. These nodes have been upgraded with i9-9900K processors. Now and all of our EOS Mainnet public API and P2P nodes are using the same i9-9900K processors to ensure transactions submitted to our API are billed at the same price that they will be executed on our block signing nodes.
We also increased the i9 servers running P2P nodes in our backup block producing data centre.
With the i9-9900K being ‘consumer’ class hardware, it takes up quite a bit of rack space compared to Xeon based servers, which we currently use for the most critical parts of our infrastructure.
Speaking of Xeon servers, we’re happy to announce that we’re in the process of deploying new Xeon Gold 6248R based servers. These heavy-duty servers have 768GB RAM and many TBs of storage space. It has the hardware nerds inside EOS Nation extremely excited. These servers will be used to continue providing history services among other things. Stay tuned to our publications for more Xeon Gold updates in the future!
Our expanding list of tools (xNation, deWeb, EOS Resource Planner, EOS Detective, Daily Snapshots, EOSIO Blocktivity statistics, BP Validator, LiquidityPortal, DSP Portal) also requires more network bandwidth to support them. So we increased network bandwidth by 8x in partnership with our provider, DNSnetworks.
What does this all mean for dApps, wallets and all EOSIO network users?
It means better infrastructure for everyone since we now offer:
- More performant EOSIO API nodes
- The ability to handle all incoming P2P requests
- Public history
- Up-to-date EOS Detective
We’re committed to investing in EOSIO infrastructure and delivering world class services to all our users. This set of changes positions us to be ready for the next phase of growth that we anticipate on our EOSIO networks.
Thank you for supporting EOS Nation. Your votes are important to us. Show your support by visiting proxy.eosnation.io!