We will use a readily available, scalable off-chain time-series database system to store microtransactions, and strategically connect it to the Ethereum blockchain. This solution will provide the performance and throughput benefits of a time series database, with all the benefits of proof of correctness of the blockchain. We are confident that we can create a system that is tightly scoped to our needs that will not be complex to design, build and deploy.
Each request made to the DService, no matter how quickly it runs, will be considered a billable event, requiring it to be logged in the form of a microtransaction. These will occur at very high volume, and the need for rapid ingestion and storage of these transactions is a critical requirement. The transaction data will always be available, to be accessed both by the regional CGN nodes to perform billing. Complete transparency in billing will allow users of the system to view and analyze all prior transactions.
The microtransactions will be recorded at a fine level of detail, i.e. per-use, but will be recorded in fiat currency so that the cost of services is expressed to each paying customer in a familiar currency. For transacting on the blockchain in Dyncoin, a scheduled process will roll up totals collectible or payable, measured in fiat currency, convert the amounts to a Dyncoin value at the time of each microtransaction, and then conduct the total amount transaction using Dyncoin, and record it on the blockchain.
In summary, we record charges and payouts in fine detail and in real time, using fiat currency. And we collect those revenues, or pay out the payables, using Dyncoin on a periodic schedule. This results in a better transaction frequency match for use with the blockchain, without sacrificing the ability to capture, collect and analyze usage data in enough detail.