You Already Know How to Use It
Market data systems have a reputation. Notoriously complicated to engineer. Difficult to deploy and operate. APIs that require weeks of study before a developer can do anything useful. This reputation is not entirely unfair — the domain is genuinely demanding, the performance requirements are extreme, and the operational stakes are high.
But a lot of the complexity in conventional market data systems is not inherent to the problem. It is the result of a design choice, usually an implicit one, to push hard problems outward: onto the developer who has to learn a proprietary API, and onto the systems engineer who has to pre-configure conflation tiers and size infrastructure for worst-case load. The platform architects solved the problems they could solve elegantly and handed the rest to the user.
Elastic MDS starts from the opposite principle. It is an autoscaling market data system: standard APIs on the outside, and a platform that adapts its own resources to the load behind them.
You Already Know How to Use It™
The APIs are industry standard: JMS for real-time low-latency push distribution, JDBC for low-latency non-push and advanced content integration. Any Java developer who has worked with a message broker or a relational database will find the programming model immediately familiar. The platform-specific behavior — live data, real-time updates, adaptive delivery — comes through those standard interfaces. No proprietary extensions to learn, no new mental model to build.
Video: S.1 — Push Distribution Code Overview. The application code is shorter than most developers expect.
A concrete example: when an application falls behind the rate at which data is arriving, the platform notices. It quietly opens additional connections to redistribute the load, reducing the pressure on any single delivery path. When the application catches up, those connections are released. The application wrote no code for this. The developer did not configure it. It happened because the platform is designed to handle it. That is what the commitment to standard APIs actually means: the hard problems stay in the platform.
Simple to use and really hard to implement well are not in tension. One requires the other. The complexity does not disappear — it moves into the platform, where it can be solved once, and solved properly.
MetaFluent holds patents on several of the mechanisms that make this possible.
Simple to engineer, simple to operate
The same principle carries through to system engineering and operations. You decide the resilience model and the resources you want to allocate. The platform handles the rest — how components find each other, how load is distributed, how the system recovers when something fails. You set the policy. The platform enforces it, and adapts as conditions change.
Observability is built on the same foundation. The entire deployment — every component, every connection, every content stream — is visible through a unified REST API. High-level state for the operations team. Fine-grained detail, down to the current value of a specific field for a specific instrument, available on demand without knowing which server holds it. One endpoint, the whole system.
Four things worth your attention
Elastic MDS tends to move the needle on four things that matter to teams running production market data infrastructure.
Lower costs. A single adaptive deployment replaces what previously required multiple dedicated systems. Resources scale with actual load. Less hardware, simpler operations.
Better resilience. No single point of failure. When something goes wrong, the impact is contained and recovery is automatic. The system is honest about what it knows and what it does not.
Easier to use. Standard APIs throughout. A developer who knows JMS or JDBC is productive on day one. You Already Know How to Use It™.
Faster time to market — and it compounds. New applications are written faster because the API is familiar. The system is easier to change — add capacity on the fly, adjust the deployment without touching application code. Further along, the platform can take on business logic directly: instead of building and operating a separate application to compute and distribute derived data, you write a line or two of SQL. The platform runs it.
What these posts cover
Each post in this series has a corresponding video showing the described behavior on a live system. We start with the immediate operational case — slow consumers, adaptive conflation, field filtering, observability — and move through architecture and resilience toward the longer-term vision of a cloud-native, auto-scaling market data and computation platform.
Start here
Video: S.2 gives a short tour of the Elastic MDS services and architecture. About five minutes. Worth it if you are new to the platform.
The goal throughout is the same as the platform itself: make something genuinely difficult look straightforward. You Already Know How to Use It™.
Related
About this blog
Elastic MDS is an autoscaling, real-time market data distribution platform developed by MetaFluent. This blog is about Elastic MDS infrastructure deployment and application development.
Further reading
