Decentralized Database Market
“Like Filecoin for databases”
Today innovative database startups struggle to get customers to switch even if their technology is faster at lower cost due to their expectation of a massive switching cost. The ddMesh network solves this problem by reducing barriers of switching resulting in : increased competition, reduced cost for customers and bringing more revenue to the most innovative providers instead of the big incumbents.
Vision
Accelerate database innovation by creating a fair and transparent market of database workloads and rewards those that can increase performance while reducing cost.
A typical companie's experience of managed database costs is that they continually go up and can particularly spike if the vendor has new owners that want contracts renegotiated. Due to each database encouraging the use of its unique features, the company's code gets locked into this vendor because it would require a huge amount of labor cost to alter all the code to target the new database technology. In our network we will prevent lock in by standardizing the API for each database type.
Stop Vendor Lock-in
All database providers, irrespective of the subnet are required to enable
automated data migration at no additional cost.
​
Try before you buy
By adding a read replica, you can test query performance on your particular workload without any line of code. And many vendors enable a free test, further lowering your barrier of switching.
RoadMap
Q1 2024
Project Inception
Q2 2024
​​Postgres subnet launch
Q3 2024
Vector subnet launch​​
Q4 2024
E2EE Document subnet launch
Q3 2025
Main net Launch
Subnets
The network is split into subnets with varying rules which focus on varying database workload specializations.
Vector Database
Primarily in production AI models, customers need to utilize a vector database for long term memory. For instance, an AI mode may map medical patient histories into an embedding space where Euclidean similarity search returns similar patients, which then could help choose the best treatment plan.
Document Database
We define a standard document store query interface inspired by the common elements of FileStore, MongoDB, DynamoDB and CosmosDB compatible API’s.
Due to the typical nature of json document queries, this subnet will be the first to release end-to-end encryption of customer data.
PostgreSQL Wire Protocol
Database providers may use any query engine and storage format which is compatible with standard SQL Select Insert and Update functions and transmits data over Postgres wire protocol. Query engines may include PostgreSQL, CockroachDB, CrateDB, YugabyteDB, etc…
ddMesh DAO
To ensure we stay aligned with our mission of increasing competition and reducing cost, our platform takes no cut of the transactions between you and your customers.
The rules of the platform are collectively governed by a DAO with votes equitably distributed between customer, provider and platform developers.
Community benefits
In addition to the direct payment for providing database resources, the network will reward additional tokens to innovators and those that provide the best customer experience.
Innovation will be measured through decentralized database benchmark competitions specific to each subnet. For instance, the vector database subnet has validators doing vector similarity search on large datasets where the database will need to do intelligent probabilistic search return an answer fast enough and validators can easily measure who got the closest answer the fastest.
Best customer experience will be measured on multiple metrics as defined by the DAO one of which is a h-index-like metric which counts how many customers have been loyal for over 6 months, 12 months, 24 months and so on. The rewards will be given out on a relative ranking within each subnet.