Db2, MQ, and WebSphere:
The Hidden Middleware Spend Nobody Owns
IBM middleware powers critical business processes, yet ownership typically falls between infrastructure, development, and application teams — meaning nobody optimises the spend. This paper provides a middleware discovery framework, maps licensing cost to business value, and presents three consolidation strategies that have reduced IBM middleware costs by 20–30% for Redress clients.
Executive Summary
IBM middleware — Db2, MQ, WebSphere Application Server, and their associated components — forms the operational backbone of enterprise IT in thousands of organisations globally. These products process transactions, route messages, host applications, and manage data at a scale that makes them indispensable. Yet in the vast majority of enterprises, IBM middleware spend falls into a governance vacuum: infrastructure teams manage the servers, development teams build the applications, and the licensing cost sits in a central IT budget that nobody optimises.
Key Findings
The Ownership Gap: Why Nobody Optimises Middleware Spend
IBM middleware cost goes unoptimised because of a structural accountability gap that exists in virtually every enterprise. Understanding this gap is the prerequisite for fixing it.
Infrastructure Manages the Servers, Not the Software. Infrastructure teams provision, patch, and maintain the physical and virtual servers running IBM middleware. They manage uptime, performance, and availability. But they do not own the software licensing decision, do not track PVU utilisation, and are not measured on middleware cost. When a new WebSphere instance is requested, infrastructure provisions a VM. Whether that instance is licensed, whether it uses the correct product edition, and whether it is still needed 3 years later are questions that infrastructure is not chartered to answer.
Development Deploys Applications, Not Licences. Application development teams select middleware platforms based on technical requirements and organisational standards. They deploy Db2 databases, configure MQ queues, and build WebSphere applications. But they do not own the licence entitlements, do not track the licensing cost of their deployments, and are not accountable for middleware spend. When a development team spins up a Db2 instance for a new project, the licensing implication is invisible to them.
Procurement Buys What’s Requested, Doesn’t Rationalise What’s Deployed. Procurement processes IBM licence purchases and S&S renewals. They negotiate pricing on transactions they are asked to execute. But procurement does not initiate middleware discovery, does not identify unused instances, and does not drive rationalisation. The S&S renewal arrives annually, procurement pays it, and the cost persists — whether the underlying middleware is still in active production use or not.
The Result: A Cost Nobody Owns. The aggregate IBM middleware cost — licences, S&S, associated infrastructure, and operational management — sits in a central IT budget as a line item that is renewed annually without review. It is not subject to the procurement scrutiny applied to new technology purchases. It is not subject to the utilisation analysis applied to cloud services. It is not subject to the rationalisation rigour applied to SaaS subscriptions. It simply persists — and grows — through inertia, annual S&S escalation, and unchecked deployment expansion.
Middleware Ownership Gap — Redress Assessment Data
than asset registers show
middleware cost owner
structured rationalisation
vs. original licence cost
The Middleware Discovery Framework
A structured middleware discovery produces the complete inventory required for any rationalisation, compliance assessment, or renewal negotiation. The framework operates across five discovery layers.
Layer 1: Entitlement Baseline. Compile a complete record of IBM middleware entitlements: Passport Advantage agreements, licence certificates, previous purchases, ULA/ELA certifications, and any acquired entity entitlements. This establishes what you own and what you’re paying S&S on — the “should be” baseline. In 81% of engagements, the entitlement baseline reveals licences for products no longer deployed and S&S payments on entitlements with no corresponding deployment.
Layer 2: Deployment Discovery. Scan all infrastructure (physical, virtual, cloud, container) for IBM middleware installations. Use ILMT as the primary discovery tool, supplemented by network-based scanning and application dependency mapping. The deployment discovery establishes the “is” state — what is actually installed and running. Compare against the entitlement baseline to identify over-deployment (installed but not licensed) and shelfware (licensed but not installed).
Layer 3: Edition & Feature Verification. IBM middleware products are available in multiple editions (Db2 Standard vs. Advanced vs. Enterprise, WebSphere Application Server Base vs. ND vs. Liberty) with significantly different PVU pricing. Verify that each deployed instance is running the edition covered by the corresponding entitlement. Edition mismatches — particularly deploying WebSphere ND features under a Base licence — are one of the most common IBM audit findings.
Layer 4: Utilisation Analysis. Measure actual utilisation of each middleware instance: connection counts, transaction volumes, message throughput, application activity, and user access. Identify instances that are installed and running but idle, lightly used, or supporting decommissioned applications. These instances are candidates for decommissioning, consolidation, or migration to lower-cost alternatives.
Layer 5: Business Value Mapping. Map each middleware instance to the business application or process it supports. Classify by criticality: mission-critical (revenue-generating, customer-facing, regulatory), business-important (internal operations, reporting, integration), and non-critical (development, test, legacy, dormant). This classification drives the rationalisation decision: mission-critical instances are optimised within IBM; non-critical instances are evaluated for migration or decommissioning.
In 91% of middleware discovery engagements, the organisation had never conducted a cross-team middleware inventory that reconciled entitlements against deployments. The most common finding is S&S payments on IBM products that are no longer deployed (pure shelfware) — averaging 12% of total IBM middleware S&S spend. This is the easiest savings to capture: terminate S&S on products you no longer use.
Db2: The Silent Cost Centre
Db2 is IBM’s relational database, deployed across thousands of enterprise environments for transaction processing, data warehousing, and application persistence. It is also the most frequently over-licensed IBM product in Redress assessments.
Edition Proliferation. Db2 is available in multiple editions (Community, Standard, Advanced Enterprise Server Edition, Enterprise Server Edition) with dramatically different PVU pricing and feature entitlements. Organisations that deployed Db2 years ago frequently run editions that were appropriate at the time of purchase but are now over-provisioned for current workloads. A Db2 Advanced Enterprise Server Edition instance supporting a departmental reporting database is paying a premium for features (partitioning, compression, high availability) that the workload does not utilise.
PostgreSQL as an Alternative. PostgreSQL has matured into a production-grade relational database with enterprise support available from EDB, Crunchy Data, and cloud-managed services (AWS RDS for PostgreSQL, Azure Database for PostgreSQL, Google Cloud SQL). For non-critical Db2 workloads — departmental databases, reporting data stores, development environments, and web application backends — PostgreSQL provides equivalent functionality at zero licence cost (or modest managed service fees). In Redress assessments, 25–35% of Db2 instances by count (not by criticality) are technically viable for PostgreSQL migration.
Cloud-Native Alternatives. Organisations migrating workloads to public cloud should evaluate cloud-native database services (AWS Aurora, Azure SQL, Google Cloud Spanner) alongside Db2 on cloud IaaS. Cloud-native services eliminate the Db2 licence cost, PVU tracking requirement, and ILMT obligation — though application compatibility, migration effort, and ongoing cloud consumption cost must be evaluated per workload.
The typical enterprise Db2 estate includes 3–5 mission-critical instances (ERP, core banking, transaction processing) that represent 60–70% of Db2 PVU consumption but only 30–40% of Db2 instance count. The remaining 60–70% of instances are non-critical workloads running high-cost Db2 editions that could be served by PostgreSQL, cloud-native databases, or lower Db2 editions at 40–60% lower cost. The rationalisation opportunity is in the long tail, not the core.
IBM MQ: Messaging Without Metrics
IBM MQ is the enterprise messaging standard, processing millions of messages per day in most large organisations. Its ubiquity makes it invisible — and its licensing cost is rarely questioned.
The Sprawl Problem. MQ queue managers proliferate organically. Every new integration project, every application refactoring, and every infrastructure migration creates queue manager instances that persist indefinitely. Development, test, and staging environments each maintain their own MQ infrastructure. The result is an MQ estate that grows steadily — and each new queue manager carries a PVU licensing obligation that is rarely tracked against the entitlement baseline.
Advanced vs. Base Licensing. IBM MQ is available in Base (standard messaging) and Advanced (extended features including Advanced Message Security, Managed File Transfer, and Telemetry) editions. Many organisations license MQ Advanced across the entire estate when only a subset of queue managers require Advanced features. Downgrading non-Advanced queue managers to MQ Base reduces per-instance PVU cost by 30–40%.
Open-Source and Cloud Alternatives. For point-to-point messaging and basic pub/sub patterns, RabbitMQ and Apache Kafka provide functionally equivalent messaging at zero licence cost. Cloud-native messaging services (AWS SQS/SNS, Azure Service Bus, Google Pub/Sub) eliminate both the licence cost and the operational overhead of managing queue manager infrastructure. However, MQ’s transactional messaging guarantees, XA transaction support, and deep integration with IBM middleware make it irreplaceable for mission-critical transactional workloads. The rationalisation opportunity is in the non-critical messaging workloads that do not require MQ’s advanced transactional guarantees.
The average enterprise MQ estate contains 40–60% more queue managers than are required for current production workloads. The surplus is attributable to abandoned project environments, dormant integration channels, duplicate development/test instances, and legacy queue managers supporting decommissioned applications. Each surplus queue manager carries a PVU licensing obligation. Decommissioning the surplus reduces MQ PVU requirements by 25–40% with zero impact on production messaging.
WebSphere Application Server: Legacy at a Premium Price
WebSphere Application Server (WAS) remains the Java application server for thousands of enterprise applications. It is also the most expensive Java application server in the market — and the one most frequently deployed in excess of what the workload requires.
The ND vs. Base vs. Liberty Decision. WebSphere is available in three editions: Network Deployment (ND), Base, and Liberty. ND provides clustering, workload management, and centralised administration at the highest PVU cost. Base provides single-server application hosting at moderate cost. Liberty provides a lightweight, modular application server at the lowest cost with the fastest start-up and smallest footprint. Many organisations deployed WAS ND 10–15 years ago when clustering was an architectural requirement. Today, container orchestration (Kubernetes, OpenShift) provides equivalent clustering capability, making WAS ND’s premium unnecessary for many workloads. Migrating from WAS ND to Liberty reduces per-instance PVU cost by 50–70%.
Open-Source Java Application Servers. Apache Tomcat, Red Hat JBoss EAP, and Eclipse Jetty provide production-grade Java application server capabilities at zero or significantly lower licence cost than WebSphere. For applications that use standard Java EE (Jakarta EE) APIs without WebSphere-specific extensions, migration to Tomcat or JBoss is technically straightforward. In Redress assessments, 30–45% of WebSphere instances run applications that do not utilise WebSphere-specific features and are technically viable for migration to open-source alternatives.
The Support Inertia Problem. Organisations continue paying WebSphere S&S on instances that have been stable for years and have not consumed an IBM support ticket in the current contract period. The S&S cost persists not because the support is needed, but because terminating S&S eliminates the right to receive patches and updates — and re-instating S&S requires back-payment of all missed fees plus IBM’s reinstatement premium. This creates a perpetual support lock-in where the fear of future need drives ongoing payment for support that is not currently consumed.
WebSphere Application Server ND at list price is approximately $30,000 per PVU core pair. A single 4-core VM running WAS ND represents $60,000 in licence value and $12,000/year in S&S. The same application running on Apache Tomcat costs $0 in licence fees. The question is whether the WebSphere-specific features that application uses are worth $12,000/year in recurring support — and in Redress assessments, the answer is “no” for 30–45% of instances.
Mapping Licensing Cost to Business Value
The most powerful rationalisation tool is a business value map that correlates each middleware instance to the business process it supports and the cost it generates. This map drives every consolidation decision.
| Classification | Criteria | Typical % of Instances | Optimisation Action |
|---|---|---|---|
| Mission-Critical | Revenue-generating, customer-facing, regulatory, SLA-bound | 15–25% | Optimise within IBM: right-size editions, consolidate VMs, negotiate S&S |
| Business-Important | Internal operations, reporting, integration, non-SLA | 30–40% | Evaluate alternatives: lower editions, Liberty, PostgreSQL, cloud-native |
| Non-Critical | Development, test, staging, dormant, legacy | 25–35% | Decommission, migrate to open source, or move to free-tier alternatives |
| Unknown / Unowned | No identified business owner, purpose unclear | 10–20% | Investigate and classify. If no owner identified within 90 days, decommission |
In the typical enterprise IBM middleware estate, 20–25% of instances generate 65–75% of business value. The remaining 75–80% of instances — non-critical, business-important but over-provisioned, or simply unknown — represent the rationalisation opportunity. The 20–30% cost reduction that Redress clients achieve comes almost entirely from this long tail, not from optimising mission-critical instances.
Three Consolidation Strategies
Three proven strategies that reduce IBM middleware costs by 20–30% without operational impact to production systems. Each strategy addresses a different dimension of the middleware cost problem.
Licence Rationalisation (Quick Wins)
Eliminate shelfware: terminate S&S on licences with no corresponding deployment. Decommission dormant instances: remove middleware from servers supporting decommissioned applications. Right-size editions: downgrade over-provisioned editions (Db2 AESE to Standard, WAS ND to Liberty, MQ Advanced to Base) on instances that do not utilise premium features. Consolidate development/test: reduce the number of non-production middleware instances through shared environments. These are operational actions that reduce cost immediately with minimal technical risk. Target timeline: 3–6 months.
Typical savings: 10–15% of total IBM middleware spend. No application changes required.
Workload Migration (Targeted Moves)
Identify non-critical and business-important middleware instances supporting workloads with viable open-source or cloud-native alternatives. Migrate eligible Db2 instances to PostgreSQL. Migrate eligible WebSphere instances to Tomcat, JBoss, or Liberty. Migrate eligible MQ queue managers to RabbitMQ or cloud messaging. Each migration eliminates the PVU licence obligation and S&S for the migrated instance. Migration is selective — mission-critical workloads remain on IBM middleware. Target timeline: 6–18 months.
Typical savings: 15–25% of total IBM middleware spend. Application testing required for each migration.
Entitlement Optimisation & Renewal Negotiation (Strategic)
Combine Strategies 1 and 2 with structured Passport Advantage renewal negotiation. Present the rationalised estate (reduced PVU requirements, decommissioned instances, migrated workloads) as the basis for renegotiating the PA agreement. Negotiate S&S reductions on the remaining estate, escalation caps for future renewals, and Flexpoint or Cloud Pak conversion terms that reflect the optimised deployment. The rationalisation work creates genuine negotiation leverage: IBM’s alternative is losing more of the estate to open-source alternatives. Target timeline: 12–18 months (aligned with PA renewal cycle).
Typical savings: 20–30% of total IBM middleware spend. Combines operational optimisation with commercial negotiation.
Common Rationalisation Traps
Six traps that prevent organisations from capturing the middleware cost reduction that the data consistently shows is available.
1. Renewing S&S Without Discovery
The most expensive habit in IBM middleware management. S&S renewals arrive annually, procurement pays them, and nobody asks whether the underlying software is still deployed or needed. A pre-renewal discovery that identifies shelfware and dormant instances should be standard practice — yet fewer than 15% of organisations conduct one.
2. Rationalising Without ILMT Validation
Organisations that reduce middleware instances without validating ILMT coverage risk creating sub-capacity compliance gaps. Every middleware change — decommissioning, consolidation, migration — must be reflected in ILMT to maintain sub-capacity eligibility. Rationalisation without ILMT governance trades cost savings for compliance exposure.
3. Treating All Middleware as Mission-Critical
Infrastructure teams resist decommissioning any middleware instance because “it might be needed.” Without a business value classification that identifies the business owner and process each instance supports, every instance is treated as critical by default. The business value mapping framework converts subjective resistance into objective classification.
4. Ignoring Edition Downgrade Opportunities
Organisations that deployed Db2 AESE or WAS ND 10+ years ago rarely revisit the edition decision. Features that were architectural requirements a decade ago may be obsolete, redundant, or replaceable with modern alternatives. An edition review that maps deployed features against edition entitlements identifies downgrade opportunities in 40–60% of instances.
5. Migrating Without Business Case
Migration from IBM middleware to open-source alternatives has both benefits (licence elimination) and costs (application testing, operational reskilling, risk). Organisations that migrate without a workload-specific business case risk under-estimating migration effort and over-estimating savings. Each migration decision should be validated by a per-workload cost-benefit analysis.
6. Negotiating PA Renewal Before Rationalising
Organisations that renegotiate their Passport Advantage agreement before conducting middleware discovery and rationalisation negotiate from an inflated baseline. IBM’s renewal is priced against the current entitlement volume. Reducing the entitlement volume through rationalisation before negotiation ensures the renewal is priced against the actual (lower) requirement — and the demonstrated ability to reduce creates negotiation leverage.
Recommendations
Seven priority actions for organisations with significant IBM middleware estates.
Assign a Middleware Cost Owner
Designate an individual or team with explicit accountability for IBM middleware cost. This owner should have visibility across infrastructure, development, and procurement; authority to initiate discovery and rationalisation; and a performance metric tied to middleware cost optimisation. Without a designated owner, middleware cost will continue to grow unchecked through inertia and organisational fragmentation.
Conduct a Cross-Estate Middleware Discovery
Map every IBM middleware instance across all infrastructure: physical, virtual, cloud, and container. Reconcile deployments against entitlements. Identify shelfware, dormant instances, edition mismatches, and unknown deployments. This discovery is the factual foundation for every rationalisation action and should be completed before any PA renewal negotiation.
Classify Every Instance by Business Value
Map each middleware instance to the business process it supports. Classify as mission-critical, business-important, non-critical, or unknown. Assign a business owner to every instance. Instances with no identified owner after 90 days should be scheduled for decommissioning. This classification drives every subsequent rationalisation decision.
Execute Quick-Win Rationalisation (Strategy 1)
Terminate S&S on shelfware. Decommission dormant instances. Right-size editions. Consolidate development/test environments. These are operational actions that deliver immediate cost reduction (10–15%) with minimal technical risk and no application changes.
Evaluate Workload Migration for Non-Critical Instances (Strategy 2)
For non-critical and business-important instances, assess migration to PostgreSQL (Db2 replacement), Tomcat/JBoss/Liberty (WebSphere replacement), and RabbitMQ/cloud messaging (MQ replacement). Produce per-workload migration business cases. Execute migrations that pass the cost-benefit test. Target 15–25% additional savings.
Align Rationalisation with PA Renewal (Strategy 3)
Time the discovery and rationalisation programme to produce results before your next Passport Advantage renewal. Present the rationalised estate as the basis for renegotiation. The demonstrated ability to reduce IBM middleware deployment creates genuine negotiation leverage that produces 20–30% total cost reduction.
Engage Specialist Advisory Support
IBM middleware rationalisation requires expertise in IBM PVU licensing, ILMT sub-capacity compliance, middleware architecture, open-source alternatives, and Passport Advantage contract negotiation. The structural ownership gap means internal teams rarely have the cross-functional expertise or mandate to drive rationalisation independently. Specialist advisory support provides the discovery capability, rationalisation framework, and negotiation expertise that consistently delivers 20–30% cost reduction.
How Redress Compliance Can Help
Redress Compliance’s IBM Practice provides end-to-end advisory support for IBM middleware discovery, rationalisation, alternative evaluation, and Passport Advantage renewal negotiation. Our team has conducted 180+ middleware rationalisation engagements with an average cost reduction of 26%.
Middleware Discovery & Rationalisation Services
- Cross-estate middleware discovery (Db2, MQ, WebSphere)
- Entitlement vs. deployment reconciliation
- Business value classification & ownership mapping
- Edition right-sizing (ND → Liberty, AESE → Standard)
- Shelfware identification & S&S termination
- Open-source migration assessment (PostgreSQL, Tomcat, RabbitMQ)
- ILMT sub-capacity compliance validation
- PVU optimisation & VM consolidation
- Passport Advantage renewal negotiation
- Ongoing middleware cost governance programme
Get In Touch
PA Renewal Approaching?
Contact us for a confidential middleware discovery and rationalisation assessment. Most organisations discover 30–40% more middleware than expected — and 20–30% cost reduction potential.
Book a Meeting
IBM middleware costs growing unchecked? Request a confidential call with our IBM Practice team.
Request a Meeting
Fill in your details and suggest times. We’ll confirm within 24 hours.
Meeting Request Sent
Thank you. Our IBM Practice team will confirm within 24 hours.
What to Expect
30-minute NDA-protected call. We’ll review your IBM middleware landscape, PA structure, and renewal timeline to scope the discovery engagement.
Based on your estate profile and our benchmark data, we’ll provide a preliminary estimate of shelfware, rationalisation potential, and expected savings.
You’ll leave with a clear plan for middleware discovery, rationalisation approach, and timeline to PA renewal — no obligation.
100% Confidential. Everything discussed is NDA-protected. We never share client data with IBM or any vendor.
No Obligation. If we can help, we’ll explain how and what it costs. If your middleware estate is already well-optimised, we’ll tell you that directly.
This document has been prepared by Redress Compliance for informational purposes. Redress Compliance is a fully independent software licensing advisory firm with zero vendor affiliations — including zero IBM partnership. We are not an IBM Business Partner and do not resell IBM products. Benchmark data is based on anonymised IBM middleware discovery and rationalisation engagements. Past results are not a guarantee of future outcomes.
© 2026 Redress Compliance. All rights reserved.