I believe its easier to add vector search to an existing cloud native database, than it is to add all the missing features required of a cloud native database to a new start up vector database. Don't believe me? Think about the following list: Networking, Security, Compliance, Role Based Access Control, Certifications, Penetration Testing, Sox Compliance, PII Certification, PCI Certification, Partitioning, Sharding, Automatic Scale out / Scale Back.... Fully Managed SaaS, DBaaS with private networking (VPC Peering, VNET Peering, Private Link, Transit Gateway, Private Services Connect) on 3 major cloud providers, AWS, GCP and Azure. But wait there is more... Replication, Automated backups, Automated Failover, Automated Recovery, Cross Region Replication for DR, Cross Cloud Replication, etc... etc... etc...) Every single item in the above list is a requirement for 90% of the Enterpise companies out there building Gen AI applications. Sooner or later the Dev team runs across road blocks by their own information security teams. How long do you think it will take these new startup vector databases to get all of the required features? My money is on MongoDB Atlas. They do stuff right, time and time again. They have ironed out all of the requirements over the past 7 years since the MongoDB IPO and exceeded expectations every single time. I expect that the Vector search capability will continue to improve with newer and better features. Click the link below to see the latest. #MongoDBAtlas #VectorSearch #RAG #generativeAI
📣 Search Nodes are now in GA for AWS, Microsoft Azure, and Google Cloud—including a memory-optimized, low CPU setup perfect for Vector Search. Dedicated infrastructure for MongoDB Atlas and Vector Search means isolated workloads, higher availability, and faster queries! Learn more: https://lnkd.in/gEQ5e6iZ
Jeff Schmitz
1mo
Really great perspective here Britton LaRoche and so true!!
not your average computer nerd
1moMongoDB Atlas is not a database. It's a data platform