[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-scaling-agentic-ai-transactional-messaging-database-en":3,"tags-scaling-agentic-ai-transactional-messaging-database-en":30,"related-lang-scaling-agentic-ai-transactional-messaging-database-en":42,"related-posts-scaling-agentic-ai-transactional-messaging-database-en":46,"series-industry-c9bc74cd-fe0d-4f7e-8d56-7a1ddbd04900":83},{"id":4,"title":5,"content":6,"summary":7,"source":8,"source_url":9,"author":10,"image_url":11,"keywords":12,"language":18,"translated_content":10,"views":19,"is_premium":20,"created_at":21,"updated_at":21,"cover_image":11,"published_at":22,"rewrite_status":23,"rewrite_error":10,"rewritten_from_id":24,"slug":25,"category":26,"related_article_id":27,"status":28,"google_indexed_at":29,"x_posted_at":10,"tweet_text":10,"title_rewritten_at":10,"title_original":10,"key_takeaways":10,"topic_cluster_id":10,"embedding":10,"is_canonical_seed":20},"c9bc74cd-fe0d-4f7e-8d56-7a1ddbd04900","Scaling Agentic AI Needs Better Data Plumbing","\u003Cp>Agentic AI looks simple in a demo: a model takes a task, calls a tool, and returns an answer. The trouble starts when that same agent has to touch payments, customer records, queues, logs, and approval systems at the same time. Oracle’s database team argues that production agentic systems need two things most prototypes skip: transactional messaging and a converged database.\u003C\u002Fp>\u003Cp>That claim matters because agentic systems are really \u003Ca href=\"\u002Fnews\u002Fdistsim-distributed-systems-on-your-laptop-en\">distributed systems\u003C\u002Fa> with a language model in the middle. Once an agent can trigger side effects, every weak spot in data consistency becomes visible. A missed event, a duplicate message, or a stale row can turn a helpful agent into an expensive support ticket generator.\u003C\u002Fp>\u003Ch2>Why demos hide the hard part\u003C\u002Fh2>\u003Cp>In a demo, the agent usually talks to one model, one tool, and one dataset. In production, the same flow often spans a vector store, a relational database, an object store, a message broker, and several APIs. That is where latency, retries, partial failures, and ordering bugs show up.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775823174135-vk5t.png\" alt=\"Scaling Agentic AI Needs Better Data Plumbing\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>Oracle’s point is simple: agentic AI magnifies the pain that already exists in distributed applications. If your system can tolerate inconsistency today, an agent can make that inconsistency more frequent and harder to diagnose. If a workflow updates inventory and sends a message on separate paths, you can end up with stock reserved in one system and no alert in another.\u003C\u002Fp>\u003Cp>The article from \u003Ca href=\"https:\u002F\u002Fblogs.oracle.com\u002Fdatabase\u002Fwhy-transactional-messaging-convergeddb-matter\" target=\"_blank\" rel=\"noopener\">Oracle Database Blog\u003C\u002Fa> argues that the database should do more than store rows. It should also coordinate events, business state, and application logic so the system has one source of truth for the transaction and the message tied to it.\u003C\u002Fp>\u003Cul>\u003Cli>Agentic systems often span 4 or more components: model, tools, data store, and queue\u003C\u002Fli>\u003Cli>Each extra hop adds failure modes such as retries, duplicates, and stale reads\u003C\u002Fli>\u003Cli>Production issues usually appear after the first 10x increase in traffic or workflow volume\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>Transactional messaging changes the failure model\u003C\u002Fh2>\u003Cp>Transactional messaging means the database commit and the message publication happen together, or not at all. That matters because many \u003Ca href=\"\u002Fnews\u002Fcursor-3-agent-interface-update-en\">agent workflows\u003C\u002Fa> depend on side effects outside the model itself. If an agent approves a refund, updates an order, and emits a notification, those actions need to stay aligned.\u003C\u002Fp>\u003Cp>Oracle’s argument is that this pattern reduces the gap between the state your application thinks it has and the state your infrastructure actually has. Instead of writing to the database and then hoping the event bus catches up, the system records the business change and the message in one transaction. That cuts down on the classic “saved in one place, missing in another” problem.\u003C\u002Fp>\u003Cblockquote>“The database is the only place where you can atomically update state and publish the event that describes that state change.” — Tim Hall, Oracle\u003C\u002Fblockquote>\u003Cp>That quote captures the practical appeal here. Agentic AI does not need more magic. It needs fewer moving parts that can disagree with each other. When the agent triggers real-world actions, atomicity stops being an academic word and becomes an incident-prevention tool.\u003C\u002Fp>\u003Cp>This also matters for auditability. If a regulated workflow uses an agent to route claims, approve spending, or update customer data, the team needs a durable record of what happened and when. A transactional system makes that record easier to trust.\u003C\u002Fp>\u003Ch2>Why a converged database matters more than another tool\u003C\u002Fh2>\u003Cp>A converged database combines multiple data models and workloads in one system instead of forcing teams to stitch together separate products for relational data, document data, vectors, and events. Oracle is pushing that idea because agentic applications often need all of those pieces at once.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775823182991-0su2.png\" alt=\"Scaling Agentic AI Needs Better Data Plumbing\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>That approach can reduce integration work, but the bigger benefit is consistency. When the same database handles operational data and AI-adjacent workloads, you reduce the number of copies, sync jobs, and handoffs in the architecture. Fewer copies mean fewer places for data drift to creep in.\u003C\u002Fp>\u003Cp>Compare that with the common stack many teams build today: PostgreSQL or MySQL for transactions, Redis for caching, Kafka for events, a vector database for retrieval, and a separate analytics store. That can work, but it also means more operational overhead and more failure points.\u003C\u002Fp>\u003Cul>\u003Cli>\u003Ca href=\"https:\u002F\u002Fwww.oracle.com\u002Fdatabase\u002F\" target=\"_blank\" rel=\"noopener\">Oracle Database\u003C\u002Fa> supports transactional and analytical workloads in one system\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fwww.postgresql.org\u002F\" target=\"_blank\" rel=\"noopener\">PostgreSQL\u003C\u002Fa> often needs extensions or external services for vector and event workflows\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fkafka.apache.org\u002F\" target=\"_blank\" rel=\"noopener\">Apache Kafka\u003C\u002Fa> is strong for event pipelines, but it is still a separate system to operate\u003C\u002Fli>\u003Cli>\u003Ca href=\"https:\u002F\u002Fredis.io\u002F\" target=\"_blank\" rel=\"noopener\">Redis\u003C\u002Fa> is fast for caching and ephemeral state, yet it does not replace a system of record\u003C\u002Fli>\u003C\u002Ful>\u003Cp>There is a trade-off, of course. A converged database can simplify the stack, but only if the team is comfortable with the vendor’s architecture and performance profile. Teams that want best-of-breed components may still prefer a modular design. The real question is whether the extra flexibility pays for the extra complexity.\u003C\u002Fp>\u003Ch2>What this means for builders shipping agentic systems\u003C\u002Fh2>\u003Cp>If you are building agents that can do real work, the architecture decision is no longer about model quality alone. It is about whether the surrounding system can preserve state, trace actions, and recover cleanly when something fails mid-flight. That is a software engineering problem as much as an AI problem.\u003C\u002Fp>\u003Cp>Oracle’s message is blunt: if your agent can create, update, or delete anything important, then your data layer needs to treat messages as first-class business events. Otherwise, you are asking the model to paper over infrastructure gaps it cannot actually fix.\u003C\u002Fp>\u003Cp>That is why the debate around transactional messaging and converged databases matters now. Agentic AI is moving from chat demos into order management, service desks, finance workflows, and internal automation. Those are exactly the places where consistency failures get expensive fast.\u003C\u002Fp>\u003Cp>For teams evaluating their stack, the takeaway is practical. Start by mapping where the agent can cause side effects, then check whether each side effect has a single atomic path from decision to durable state. If the answer is no, the next outage may look like an AI problem, but the root cause will probably be data plumbing.\u003C\u002Fp>\u003Cp>The likely near-term pattern is simple: teams that keep agents on top of fragile, multi-system workflows will spend more time debugging than shipping. Teams that tighten the transaction boundary around the agent’s actions will spend less time reconciling state and more time improving the actual product. The question is not whether agents can act, but whether your database can keep up when they do.\u003C\u002Fp>","Agentic AI breaks at scale when data and events drift. Oracle argues transactional messaging and a converged database can keep systems consistent.","blogs.oracle.com","https:\u002F\u002Fblogs.oracle.com\u002Fdatabase\u002Fwhy-transactional-messaging-convergeddb-matter",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1775823174135-vk5t.png",[13,14,15,16,17],"agentic ai","transactional messaging","converged database","oracle database","distributed systems","en",0,false,"2026-04-10T12:12:34.272962+00:00","2026-04-10T12:12:34.199+00:00","done","32d1b78a-8d47-42a6-8815-0abba46194c0","scaling-agentic-ai-transactional-messaging-database-en","industry","1903476f-c939-4518-be7e-67f3562de087","published","2026-04-11T09:00:07.465+00:00",[31,33,35,37,39],{"name":16,"slug":32},"oracle-database",{"name":17,"slug":34},"distributed-systems",{"name":14,"slug":36},"transactional-messaging",{"name":15,"slug":38},"converged-database",{"name":40,"slug":41},"agentic AI","agentic-ai",{"id":27,"slug":43,"title":44,"language":45},"scaling-agentic-ai-transactional-messaging-database-zh","代理型 AI 要先補資料管線","zh",[47,53,59,65,71,77],{"id":48,"slug":49,"title":50,"cover_image":51,"image_url":51,"created_at":52,"category":26},"6ff3920d-c8ea-4cf3-8543-9cf9efc3fe36","circles-agent-stack-targets-machine-speed-payments-en","Circle’s Agent Stack targets machine-speed payments","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778871659638-hur1.png","2026-05-15T19:00:44.756112+00:00",{"id":54,"slug":55,"title":56,"cover_image":57,"image_url":57,"created_at":58,"category":26},"1270e2f4-6f3b-4772-9075-87c54b07a8d1","iren-signs-nvidia-ai-infrastructure-pact-en","IREN signs Nvidia AI infrastructure pact","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778871059665-3vhi.png","2026-05-15T18:50:38.162691+00:00",{"id":60,"slug":61,"title":62,"cover_image":63,"image_url":63,"created_at":64,"category":26},"b308c85e-ee9c-4de6-b702-dfad6d8da36f","circle-agent-stack-ai-payments-en","Circle launches Agent Stack for AI payments","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778870450891-zv1j.png","2026-05-15T18:40:31.462625+00:00",{"id":66,"slug":67,"title":68,"cover_image":69,"image_url":69,"created_at":70,"category":26},"f7028083-46ba-493b-a3db-dd6616a8c21f","why-nebius-ai-pivot-is-more-real-than-hype-en","Why Nebius’s AI Pivot Is More Real Than Hype","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778823055711-tbfv.png","2026-05-15T05:30:26.829489+00:00",{"id":72,"slug":73,"title":74,"cover_image":75,"image_url":75,"created_at":76,"category":26},"b63692ed-db6a-4dbd-b771-e1babdc94af7","nvidia-backs-corning-factories-with-billions-en","Nvidia backs Corning factories with billions","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778822444685-tvx6.png","2026-05-15T05:20:28.914908+00:00",{"id":78,"slug":79,"title":80,"cover_image":81,"image_url":81,"created_at":82,"category":26},"26ab4480-2476-4ec7-b43a-5d46def6487e","why-anthropic-gates-foundation-ai-public-goods-en","Why Anthropic and the Gates Foundation should fund AI public goods","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1778796645685-wbw0.png","2026-05-14T22:10:22.60302+00:00",[84,89,94,99,104,109,114,119,124,129],{"id":85,"slug":86,"title":87,"created_at":88},"d35a1bd9-e709-412e-a2df-392df1dc572a","ai-impact-2026-developments-market-en","AI's Impact in 2026: Key Developments and Market Shifts","2026-03-25T16:20:33.205823+00:00",{"id":90,"slug":91,"title":92,"created_at":93},"5ed27921-5fd6-492e-8c59-78393bf37710","trumps-ai-legislative-framework-en","Trump's AI Legislative Framework: What's Inside?","2026-03-25T16:22:20.005325+00:00",{"id":95,"slug":96,"title":97,"created_at":98},"e454a642-f03c-4794-b185-5f651aebbaca","nvidia-gtc-2026-key-highlights-innovations-en","NVIDIA GTC 2026: Key Highlights and Innovations","2026-03-25T16:22:47.882615+00:00",{"id":100,"slug":101,"title":102,"created_at":103},"0ebb5b16-774a-4922-945d-5f2ce1df5a6d","claude-usage-diversifies-learning-curves-en","Claude Usage Diversifies, Learning Curves Emerge","2026-03-25T16:25:50.770376+00:00",{"id":105,"slug":106,"title":107,"created_at":108},"69934e86-2fc5-4280-8223-7b917a48ace8","openclaw-ai-commoditization-concerns-en","OpenClaw's Rise Raises Concerns of AI Model Commoditization","2026-03-25T16:26:30.582047+00:00",{"id":110,"slug":111,"title":112,"created_at":113},"b4b2575b-2ac8-46b2-b90e-ab1d7c060797","google-gemini-ai-rollout-2026-en","Google's Gemini AI Rollout Extended to 2026","2026-03-25T16:28:14.808842+00:00",{"id":115,"slug":116,"title":117,"created_at":118},"6e18bc65-42ae-4ad0-b564-67d7f66b979e","meta-llama4-fabricated-results-scandal-en","Meta's Llama 4 Scandal: Fabricated AI Test Results Unveiled","2026-03-25T16:29:15.482836+00:00",{"id":120,"slug":121,"title":122,"created_at":123},"bf888e9d-08be-4f47-996c-7b24b5ab3500","accenture-mistral-ai-deployment-en","Accenture and Mistral AI Team Up for AI Deployment","2026-03-25T16:31:01.894655+00:00",{"id":125,"slug":126,"title":127,"created_at":128},"5382b536-fad2-49c6-ac85-9eb2bae49f35","mistral-ai-high-stakes-2026-en","Mistral AI: Facing High Stakes in 2026","2026-03-25T16:31:39.941974+00:00",{"id":130,"slug":131,"title":132,"created_at":133},"9da3d2d6-b669-4971-ba1d-17fdb3548ed5","cursors-meteoric-rise-pressures-en","Cursor's Meteoric Rise Faces Industry Pressures","2026-03-25T16:32:21.899217+00:00"]