At Percona, our mission has all the time been that will help you succeed with open supply databases. We do this by listening to the neighborhood, understanding the challenges you face, and constructing the options you want. Now, after a complete evaluate of market tendencies and direct suggestions from our prospects and the MySQL neighborhood, we’re excited to announce our plans for 2 main new initiatives for the MySQL ecosystem: Vector Search & Indexing and a devoted MySQL Binlog Server.
This isn’t a selection between two completely different paths; it’s a dedication to 2 complementary pillars of our technique. We consider that to be the most effective associate for you, we should be each a Trusted associate who solves at this time’s hardest issues and an Innovator who builds for tomorrow’s alternatives. These initiatives ship on each of these guarantees.
Vector Assist: Embracing the AI-Powered Future
The rise of AI and machine studying is reshaping how purposes are constructed. The demand for vector search capabilities is not a distinct segment requirement; it’s a strategic necessity. We’ve seen this clearly out there: a latest focused neighborhood ballot confirmed that 68% of enterprise leaders view built-in AI/vector workloads as a beneficial future requirement, and a LinkedIn ballot by our founder, Peter Zaitsev, confirmed that 73% of the neighborhood believes MySQL wants this performance.
Extra importantly, we’ve acquired robust curiosity from strategic enterprise accounts and have seen the aggressive danger firsthand as firms take into account the way forward for the MySQL ecosystem.
Why Our Vector Answer Will Be Higher
Whereas a number of vector options exist within the MySQL ecosystem, every comes with important trade-offs. Our aim is to ship an answer that’s superior for our customers by being a local, absolutely open supply, and true drop-in substitute for MySQL.
- In contrast to DBaaS options, which lock you into a selected proprietary cloud platform, our answer will probably be absolutely open supply, supplying you with the liberty to run anyplace.
- In contrast to MySQL-compatible engines, which require a fancy migration, our function will probably be a easy, drop-in enhancement to the Percona Server for MySQL you already belief.
- In contrast to primary plugins or connectors, which perpetuate the “two-system drawback” we purpose to unravel, our native answer will present the deep integration and transactional (ACID) ensures that enterprises want for mission-critical purposes.
Our plan is to ship an answer that eliminates the complexity, price, and information consistency complications of managing a separate, specialised vector database. The preliminary model will concentrate on delivering native vector search and indexing, constructed on three core capabilities you want:
- Performant Search and Indexing: Create devoted indexes on vector information and run quick, approximate nearest neighbor (ANN) searches.
- Seamless Integration with Normal SQL: Mix vector searches with the
WHEREclauses and enterprise logic you employ on daily basis, all in a single question. - Full Transactional (ACID) Consistency: Guarantee your vector information is as dependable and constant as the remainder of your transactional information.
Binlog Server: Reinforcing Operational Excellence at Scale
Whereas we construct for the long run, we stay deeply dedicated to fixing the fast, real-world challenges our enterprise prospects face at this time. For these operating MySQL at scale, managing catastrophe restoration and distributed architectures stays a major supply of operational ache.
The demand for a greater answer right here will not be speculative; it’s a direct and pressing request from large-scale enterprise customers operating complicated, business-critical environments. They’ve informed us they want a extra strong and environment friendly solution to handle their binary logs for Level-in-Time Restoration (PITR) and to scale their replication topologies with out overloading their main servers.
The Worth We Plan to Construct:
This isn’t nearly archiving logs. Our imaginative and prescient is to create a foundational piece of infrastructure that makes your complete database surroundings extra resilient and environment friendly.
The Binlog Server will probably be an energetic, on-line service centered on the core “Job to be Performed”: enabling dependable, automated Level-in-Time Restoration. Key capabilities will embody appearing as a GTID-based dwell replication supply, automated GTID state administration, and enabling exact PITR workflows by timestamp or GTID set.
The place We Are At the moment & How You Can Assist
We’re firmly dedicated to constructing each of those high-impact options. Our fast focus is on the preliminary functionality validation course of. We’re working to validate a hyper-focused scope for every initiative to make sure we are able to ship significant worth to you as shortly as potential.
That is the place you are available.
To make sure we construct the best options, we have to validate our plans with the individuals who will use them on daily basis. We’re in search of customers, builders, and DBAs to take part in our validation course of. That is your alternative to offer important suggestions on our proposed preliminary scope and assist form the way forward for these important new instruments for the MySQL ecosystem.
If you’re constructing (or have plans to construct) AI purposes on MySQL or in case you are dealing with the challenges of managing catastrophe restoration at scale, we need to hear from you! Please attain out to our MySQL Product Supervisor to take part on this important subsequent step. Ebook a 1-1 assembly right here. Alternatively, please add your feedback beneath to take part within the dialog.
Thanks for being a part of the Percona neighborhood. Collectively, we are going to proceed to construct the way forward for MySQL.
