Middleware

The unseen connective tissue of enterprise technology — the software layer that lets your systems talk to each other.

The global middleware market exceeded $45 billion in 2025 and is projected to reach $70 billion by 2029, driven by the explosive growth of APIs, event-driven architectures, and the need to connect an ever-expanding ecosystem of SaaS applications, legacy systems, and AI services. For most enterprises, middleware is the single largest category of software infrastructure spending after cloud compute itself — and one of the least understood by business leaders. When middleware fails, everything stops: orders cannot be processed, payments cannot be settled, and customers cannot be served.

What It Is

Middleware is software that sits between an operating system and the applications that run on it, providing common services and capabilities that allow different systems to communicate and work together. It is the plumbing and wiring of enterprise technology — invisible when it works, catastrophic when it does not. If your customer relationship management (CRM) system needs to send order data to your enterprise resource planning (ERP) system, which in turn needs to trigger a fulfilment process in your warehouse management system, middleware is the layer that connects these systems, translates between their different data formats, and ensures messages are delivered reliably.

The concept has been part of enterprise computing since the 1990s, when technologies like CORBA and Message-Oriented Middleware emerged to connect disparate systems. Modern middleware has evolved into several distinct categories: message brokers (such as Apache Kafka and RabbitMQ) that enable asynchronous communication between services; API gateways (such as Kong and Apigee) that manage and secure application programming interfaces; enterprise service buses (ESBs) that provide transformation and routing; and integration platform as a service (iPaaS) offerings that provide cloud-native connectivity between SaaS applications. Each category addresses a different integration pattern, and most large enterprises use a combination of several.

A useful analogy: if enterprise applications are the rooms in an office building, middleware is the corridor system, elevators, ventilation, and electrical wiring that connect them and make the building functional. You never visit the building to admire the wiring, but without it, every room is isolated and unusable. Middleware ensures that when a customer places an order on your website, the inventory system is updated, the payment is processed, the shipping label is generated, and the confirmation email is sent — all orchestrated across multiple independent systems.

Why It Matters for Business

Middleware directly determines the agility of your technology estate. Organisations with well-designed middleware can integrate a new SaaS tool, partner API, or acquired company's systems in days or weeks. Those without it face months of custom integration work for each new connection. The average enterprise uses over 370 SaaS applications, and each one needs to connect to the others. Without middleware, the number of point-to-point integrations grows factorially — connecting 10 systems requires 45 individual integrations; connecting 20 systems requires 190. Middleware reduces this to a manageable hub-and-spoke or publish-subscribe model.

The cost of middleware failure is severe. When a message broker goes down, orders may be lost, payments may fail to process, and customer data may become inconsistent across systems. Major outages attributed to middleware failures include the 2023 T-Mobile billing system crash (caused by an overloaded integration layer) that prevented customers from making payments for 48 hours, and the 2022 Southwest Airlines operational meltdown partly attributed to middleware failures connecting scheduling and crew management systems. Both incidents cost over $100 million in direct losses and reputational damage.

Middleware also plays a critical role in regulatory compliance. Data residency requirements (such as those under GDPR and PDPA), audit trails for financial transactions, and real-time monitoring for fraud all depend on middleware that can route, log, and transform data according to policy. Organisations subject to regulations such as SOX, HIPAA, or PSD2 rely on middleware to enforce data handling rules consistently across all connected systems.

Enterprise Use Cases

Walmart processes over 1 billion messages per day through its Apache Kafka middleware layer, connecting point-of-sale systems, inventory management, supplier portals, and e-commerce platforms. During Black Friday, the system handles peak throughput of over 10 million events per second, ensuring that inventory levels are synchronised across 4,700 US stores and the online store in real time. If a customer buys the last television in a store, the online inventory updates within seconds to reflect the new stock level, preventing overselling and the resulting customer dissatisfaction.

Goldman Sachs uses an enterprise middleware platform to connect over 5,000 internal applications and external market data feeds. The system processes millions of financial messages per day across equities, fixed income, derivatives, and foreign exchange, with latency requirements measured in microseconds. The middleware layer handles message transformation between different financial protocols (FIX, SWIFT, proprietary formats), ensuring that a trade initiated in one system is correctly understood by every downstream system. Royal Bank of Scotland migrated over 200 legacy integrations to a modern API-based middleware platform, reducing the time to onboard new fintech partners from 6 months to 3 weeks.

In healthcare, Kaiser Permanente uses middleware to connect electronic health records across 39 hospitals and 723 medical offices, ensuring that a patient's records are available to any clinician at any location within seconds. The system processes over 50 million clinical messages daily, including lab results, prescription orders, and imaging reports.

Vendor Landscape

The middleware market spans traditional and cloud-native providers. For enterprise integration, MuleSoft (acquired by Salesforce for $6.5 billion in 2018) is the market leader in iPaaS, with pricing typically ranging from $100,000 to over $1 million annually depending on connection volume. Boomi (spun out of Dell) and Workato are strong competitors, with Workato particularly popular for its "recipe" approach that business users can configure without deep technical skills. For message streaming, Confluent (the commercial entity behind Apache Kafka) dominates with enterprise contracts starting at $150,000 per year, whilst Amazon Kinesis, Azure Event Hubs, and Google Pub/Sub offer cloud-native alternatives. API management leaders include Apigee (Google), AWS API Gateway, Kong, and Tyk. Traditional vendors such as IBM (App Connect, MQ) and TIBCO continue to serve large enterprises with legacy middleware estates, though they are gradually losing ground to cloud-native alternatives.

Strategic Considerations

Treat middleware as a strategic platform investment, not a tactical integration tool. Selecting middleware based solely on the immediate need to connect two systems is a common mistake that leads to a fragmented landscape of incompatible integration tools. Establish an integration strategy and platform standard before proliferating middleware solutions. Evaluate total cost of ownership carefully — the licence cost of middleware is typically 20-30% of the total cost, with the remainder consumed by development, maintenance, and operational support. For organisations with extensive legacy systems, consider an incremental modernisation approach: use modern API-based middleware to wrap legacy systems, exposing their capabilities as APIs that newer services can consume without requiring immediate legacy replacement. Invest in middleware monitoring and observability — integration failures are among the hardest to diagnose and the most damaging when they occur. Finally, consider whether your middleware strategy needs to support event-driven architecture (where systems react to events in real time) or whether simpler request-response patterns suffice; the choice has significant implications for vendor selection and organisational readiness.