#cloudgraph

CloudGraph currently runs on both local and remote MySQL servers.

We have just begun implementing MySQL as a backend for CloudGraph. The new MySQL backend will perform slower than our current in-house-developed backend, but will allow a wider variety of deployment scenarios. Until the clustered version of CloudGraph is production ready, MySQL will provide better cross-server scalability and give developers a chance get started in a more familiar environment. We expect to integrate other datastores like Microsoft SQL (and Azure) over time. Our initial proof of concept for Webling was built on a MySQL backend - and we haven't gone mad by reimplementing it. The approach simply allows us to separate the business logic from the database layer and use MySQL as a substitute for disk-based files.

Today we've released a complete redesign of www.cloudgraph.com. This upgrade also reveals CloudGraph's new logo. We hope you like it!

CloudGraph will give web hosting service providers the ability to supply fast and intuitive .NET graph databases to their customers. CloudGraph has a feature-rich API that will allow web hosting service providers to create graph databases on demand and enable customer isolation, with fine-grain control over server ressources.

The size of one block is equivalent to a raw node. Min. throughput represents an empty cache (cold start). Max throughput represents the database-file already in cache (running server). Note that multi-threaded sequential read and write are not actually sequential, because multiple threads are trying to read and write simultaneously. Page size has a large effect on sequential read and write performance.

For the last couple of months we've been looking into using GPU acceleration in our graph database. By utilizing the numerous thread processors in modern graphic cards, we hope to exceed the traditional boundaries of CPU concurrency. We think it's possible for us to implement an extremely fast traversal logic using GPU memory as read-only hot-cache and look forward to further testing the concept. More details to follow...