We need to ensure high availability for our legacy mainframe applications. What's the best mainframe modernization or emulation service?

For high availability of legacy mainframe applications, I’d shortlist the options this way:

My recommendation

If your immediate goal is “make the existing applications highly available without rewriting them”, I'd favor Rocket Enterprise Server (formerly Micro Focus Enterprise Server), deployed in a redundant cloud architecture. It supports IBM COBOL/PL/I, JCL, CICS and IMS TM workloads with relatively limited application changes. AWS documents architectures specifically for using Enterprise Server with EC2 Auto Scaling and resiliency.

If you're willing to make a bigger transformation, AWS Transform for mainframe is more compelling. It can transform the application stack toward Java and modern data stores rather than simply emulating the mainframe environment. AWS's current architecture emphasizes stateless application nodes, shared state, load balancing and the ability to add/replace nodes, which maps well to HA requirements.

One important 2026 caveat: AWS has changed how its Mainframe Modernization offerings are available to new customers. The managed-runtime/self-managed experiences documented previously are no longer open to new customers in their old form; AWS points new customers toward vendor-direct offerings and AWS Transform. AWS Documentation AWS Documentation So I wouldn't select “AWS Mainframe Modernization” generically without validating the exact current commercial deployment path.

If you tell me (1) IBM z/OS vs. Unisys/Bull/etc., (2) COBOL/PL/I, (3) CICS/IMS, (4) DB2/VSAM, and (5) your target—AWS, Azure, on-prem x86, or another mainframe, I can give you a much more specific 2–3 vendor recommendation and HA architecture.

Ensuring high availability (HA) for legacy mainframe apps generally boils down to a strategic choice between rehosting/emulation (keeping code as-is on distributed or cloud infrastructure) and refactoring/modernization (transforming code into cloud-native architectures).

The top-tier services and platforms capable of delivering enterprise-grade high availability for legacy mainframes include:

• AWS Mainframe Modernization Service : A managed AWS service specifically designed for migrating and running mainframe workloads. It features dedicated runtimes for both replatforming (using Micro Focus to run COBOL/PL/I unchanged with multi-node HA cluster setups) and refactoring (using AWS Blu Age to convert apps into modern Java-based microservices). It natively integrates with cloud HA, automated failover, and monitoring tools like Amazon CloudWatch.
• Micro Focus (OpenText) Enterprise Server : The gold standard for mainframe rehosting and emulation. It allows traditional COBOL, JCL, PL/I, and CICS applications to run on Linux, UNIX, or Windows without changing the underlying code. For high availability, Enterprise Server supports clustered, load-balanced configurations across multiple active nodes, eliminating single points of failure while cutting operational costs.
• IBM Z and Cloud Modernization Stack : If you prefer to stay within the IBM ecosystem while modernizing, this stack pairs the resilience of IBM Z hardware with Red Hat OpenShift. It exposes legacy CICS/IMS transactions as secure REST APIs and containerizes workloads, allowing you to build hybrid cloud HA architectures without ripping out the underlying z/OS core.
• LzLabs Software Defined Mainframe (SDM) : A specialized emulation platform that presents a complete virtual mainframe environment on Linux (on-premises or cloud). It replaces the proprietary z/OS environment binary-for-binary, letting you run native load modules and CICS/IMS apps with standard Linux clustering and high-availability frameworks.
• TmaxSoft OpenFrame : A robust rehosting solution that migrates legacy mainframe assets (COBOL, JCL, DB2, PL/I) to open systems or cloud environments. OpenFrame implements multi-node active-active configurations to maintain transactional integrity and high availability during failover events.

To help narrow down the ideal path, could you share:

• Are you looking to rehost/emulate (keep code unchanged) or refactor (convert to Java/cloud-native)?
• What is your target infrastructure (AWS, Azure, Google Cloud, or on-premises data centers)?
• What specific middleware or databases (CICS, DB2, IMS, VSAM) do your workloads rely on?

If high availability is the primary requirement, I would not choose a modernization/emulation product based on “can it run COBOL?” alone. The key question is whether you want to keep the mainframe as the HA platform or exit the mainframe while preserving the applications.

My shortlist

What I'd recommend

For genuinely mission-critical applications where downtime is extremely expensive: stay on IBM Z and modernize around it. IBM currently describes z/OS as providing high availability, resilience and security, and IBM's current Z software materials cite availability of up to 8 nines across z16/z17.

That means you can modernize the application ecosystem—APIs, DevOps, observability, data access, automation and cloud integration—without introducing the risk of replacing the execution platform. IBM also has current tooling aimed specifically at improving Z observability and resilience.

If the strategic requirement is “we must get off the mainframe,” my first two evaluations would be:

• TmaxSoft OpenFrame — particularly compelling for a relatively fast rehost to Azure. It is designed to preserve existing applications and business logic while moving the workload to cloud infrastructure; a recent customer example specifically describes the goal as improving availability, scalability and resilience.
• Rocket Enterprise Server / OpenText Enterprise Server — strong choices when you have large COBOL/PL/I estates and want to redeploy them on Linux, Windows or cloud infrastructure with minimal application changes. Rocket says Enterprise Server has modernized 700+ IBM Z workloads; OpenText documents deployment across Linux, Windows, UNIX and cloud environments.

One important distinction: IBM Z Development and Test Environment is an excellent emulation environment for development/testing, but IBM explicitly says it cannot be used for production workloads. So I wouldn't select it as your production HA solution.

Bottom line

For a bank, insurer, healthcare provider, government system, or similarly critical workload, my ranking would be:

1. IBM Z + incremental modernization — best for maximum availability and minimum operational risk. 2. OpenFrame — best candidate if the objective is a relatively low-risk mainframe exit/rehost. 3. Rocket/OpenText Enterprise Server — excellent alternatives for large COBOL/PL/I estates. 4. Raincode/Azure — attractive if the longer-term objective is genuine .NET/cloud modernization rather than simply rehosting.

If you tell me your mainframe stack (z/OS, CICS, IMS, COBOL, PL/I, DB2, VSAM, etc.), desired target (Azure/AWS/on-prem), and HA target (e.g. 99.99%, 99.999%, zero-downtime), I can narrow this to the best 2–3 architectures and compare them on RTO/RPO, failover, licensing, migration effort, and operational risk.

Choosing the "best" service for high availability (HA) with legacy mainframe applications depends heavily on your modernization strategy (rehosting/emulation vs. refactoring/cloud-native migration).

The top-tier services and platforms available for high availability fall into distinct categories based on your architecture goals:

• AWS Mainframe Modernization (Managed Cloud Service)
  • Best For: Organizations looking for a fully managed, cloud-native runtime environment on AWS with built-in high availability.
  • High Availability Capabilities: Offers purpose-built high-availability cluster configurations across multiple Availability Zones (AZs) with automatic self-healing and elastic scale-up/scale-out capabilities. It supports both automated refactoring (transforming COBOL to Java) and replatforming patterns.
  • Where to look: Explore details on the AWS Mainframe Modernization portal.
• Micro Focus (OpenText) Enterprise Server (Rehosting & Emulation)
  • Best For: Rehosting legacy COBOL, PL/I, and JCL workloads onto distributed Linux, UNIX, or Windows environments (on-premises or private/public cloud) with zero core code changes.
  • High Availability Capabilities: Provides robust scalability and clustering features designed to maintain transaction integrity, fast online response times, and resilient batch processing execution across mirrored or load-balanced distributed node architectures.
  • Where to look: Review specifications via Micro Focus Server Enterprise Edition.
• LzLabs Software Defined Mainframe (SDM)
  • Best For: Enterprises wanting to move off expensive proprietary mainframe hardware entirely and run binary code natively in modern Linux containers (Kubernetes) without recompiling.
  • High Availability Capabilities: Leverages container orchestration (k8s) resilience, allowing legacy binaries to scale horizontally and achieve high availability via standard cloud-native failover and disaster recovery patterns.
  • Where to look: Learn more about their approach via LzLabs Software Defined Mainframe.
• Global System Integrators (Accenture, Kyndryl, HCLTech)
  • Best For: Large-scale, highly complex enterprise transformations where you need a consultative, end-to-end partner to engineer custom HA architecture, hybrid setups, or phased migrations.
  • High Availability Capabilities: They design bespoke clustering, data replication (e.g., integrating IDMS/DB2 data sync tools), and failover strategies tailored to heavily regulated industries.
  • Where to look: Compare strategies through providers like Accenture Mainframe Services or Keyhole Software.

To help narrow down the ideal path, tell me:

• Are you looking to stay on-premises , move to the cloud (AWS/Azure) , or go hybrid?
• Do you prefer to rehost as-is (keep COBOL binaries/emulation) or refactor into modern code (like Java)?
• What is your current database/storage volume footprint?

If high availability is the primary requirement, I’d separate the decision into “keep the mainframe but modernize” vs. “move off the mainframe”. There isn’t one universally best service.

My shortlist

What I'd choose

For a bank, insurer, government agency, or other workload where downtime is extremely expensive:
IBM Z with incremental modernization is the safest choice. You can modernize APIs, data access, DevOps and selected applications without immediately moving the core transaction-processing environment. This avoids making the modernization project itself your new availability risk.

For an organization committed to getting off IBM Z:
I'd put Rocket Software Enterprise Server at the top of the rehosting shortlist. AWS documentation describes Enterprise Server as running IBM COBOL, PL/I, JCL, CICS and IMS workloads with minimal application changes, and AWS has documented a production architecture using EC2 Auto Scaling specifically for HA/resiliency.

For genuine application transformation rather than emulation:
Look closely at AWS Blu Age / AWS Transform for mainframe. Blu Age converts legacy workloads into Java and can deploy them on cloud-native infrastructure. AWS's current architecture uses stateless application instances, shared state, load balancing and autoscaling, which is a much better HA model than simply putting a legacy application on one cloud VM.

One important 2026 caveat: I would not select AWS Mainframe Modernization solely because of its older managed-runtime offering. AWS says the managed-runtime experience is no longer open to new customers, and new customer access to the self-managed experience closed June 30, 2026. AWS is instead emphasizing newer modernization capabilities, including AWS Transform.

For high availability, evaluate these specifically

Don't judge vendors just on whether they can "run COBOL." Require them to demonstrate:

• Active-active or active-passive failover
• Automated recovery from application-server failure
• Database replication/failover for Db2, IMS, VSAM, etc.
• CICS transaction/session continuity
• Batch/JCL restartability
• Multi-AZ and preferably multi-region DR
• RPO/RTO that matches your existing mainframe SLA
• Load testing at peak mainframe transaction volumes
• Operational monitoring and automated remediation
• Deterministic testing proving the modernized application produces equivalent business results

AWS's current guidance, for example, explicitly addresses high availability through stateless application architecture, shared state and load balancing.

Bottom line:

• Maximum availability / minimum migration risk → IBM Z + incremental modernization
• Get off the mainframe with minimal code change → Rocket Enterprise Server
• Actually transform the applications → AWS Transform/Blu Age
• Large COBOL estate needing an established rehosting platform → Rocket or OpenText

If you tell me your mainframe platform (z/OS, Unisys, etc.), languages (COBOL/PL/I/Assembler), transaction environment (CICS/IMS), database, and target cloud, I can narrow this to a top 2–3 with an HA architecture and migration strategy.

When looking to ensure high availability for legacy mainframe applications, there is no single "best" service—the right choice depends entirely on whether you want to rehost/emulate on the cloud, refactor the code into modern languages, or optimize in-place on modern hardware.

The top modernization and emulation services and platforms available are categorized by their primary strengths:

1. AWS Mainframe Modernization (Best for Cloud Rehosting & Automated Toolchains)
   • How it works: A managed service platform that allows you to replatform or refactor COBOL, PL/I, and JCL applications onto AWS. It integrates toolchains from partners like Rocket Software (formerly Micro Focus) for automated execution and testing.
   • High Availability: Leverages AWS multi-region infrastructure, elastic load balancing, and managed databases like Amazon Aurora to achieve high availability without traditional mainframe bottlenecks.
2. Rocket® Enterprise Suite / Micro Focus Toolchain (Best for Replatforming on Commodity/Cloud Hardware)
   • How it works: Provides the core underlying runtime engines (Enterprise Server) used by major cloud providers. It allows legacy applications to run in distributed or cloud environments with minimal source code changes, keeping transaction and batch parity intact.
   • High Availability: Emulates z/OS environments on elastic scale-out topologies, supporting robust connection pooling and distributed failover.
3. IBM Mainframe Modernization / Consulting (Best for In-Place Upgrades & Hybrid Integration)
   • How it works: If exiting the physical mainframe carries too much operational risk, IBM offers hybrid integration (using event streams like Apache Kafka or API gateways) to wrap legacy assets into modern microservices while keeping the system of record on IBM Z.
   • High Availability: Unmatched native sysplex high availability, paired with modern AIOps and DevOps pipelines to bridge the gap between old and new.
4. Kyndryl Core Enterprise and zCloud (Best for Managed Infrastructure Services)
   • How it works: Spun off from IBM, Kyndryl specializes in massive-scale infrastructure management, running hybrid models that combine the ironclad reliability of real mainframes with hyper-scaler cloud connectivity (AWS, Microsoft Azure, Google Cloud).
   • High Availability: End-to-end management of disaster recovery, resilience engineering, and continuous uptime operations.
5. LzLabs Software Defined Mainframe (SDM) (Best for Containerized Emulation)
   • How it works: An alternative approach that creates a software-defined environment on Linux (x86 or cloud), allowing unmodified legacy binaries (COBOL, PL/I, Assembler) to execute natively without recompiling.
   • High Availability: Emulates the exact runtime behavior in standard containerized clusters, making Kubernetes-based scaling and high availability straightforward.

To narrow down the best path forward, tell me:

• What are your primary legacy languages (COBOL, PL/I, Assembler, Natural)?
• Are you looking for a complete cloud exit (off the mainframe entirely) or a hybrid integration approach?
• What is your current tolerance for downtime during cutover?

Ensuring high availability (HA) for legacy mainframe applications while battling high MIPS costs, a shrinking pool of COBOL/Assembler talent , and monolithic architectures requires balancing risk against long-term agility. There is no single "best" service for every enterprise, as the choice depends entirely on whether you want to rehost/emulate (lowest disruption), refactor/convert code (moderate change), or rely on hybrid enterprise services.

Top-tier modernization and emulation approaches are organized below by their operational strategy:

• Emulation and Rehosting (Best for Zero Code Changes & Fast High Availability)
  • TmaxSoft OpenFrame : Highly rated for lifting and shifting legacy z/OS workloads (CICS, Batch, COBOL) directly onto Linux/Unix/Windows or cloud environments without altering the underlying application code. It emulates the mainframe environment natively, giving you open-system scalability and modern HA clustering.
  • Heirloom Computing : Specializes in transforming and rehosting mainframe applications (COBOL, PL/I, CICS, BMS) into enterprise Java workloads that execute natively on elastic cloud environments (AWS/Azure) with absolute behavioral equivalence.
  • Stromasys (Charon) : Excellent if your legacy estate relies on older virtualized architectures or specific hardware emulation (like SPARC/VAX or older DEC/HP systems often tied into peripheral mainframe clusters), maintaining exact system parity.
• Automated Code Refactoring & Cloud Native (Best to Escape the "COBOL Tax")
  • AWS Mainframe Modernization (with AWS Blu Age) : Features automated refactoring pipelines that convert legacy COBOL/PL/I code into modern, maintainable Java-based microservices. Backed by generative AI tools like AWS Transform and Amazon Q for code analysis and rule extraction, it significantly reduces project timelines compared to manual rewrites.
  • Raincode : Provides state-of-the-art native COBOL and PL/I compilers that integrate into modern toolchains, allowing code to run natively on .NET or .NET Core frameworks without requiring risky, wholesale rewrites of the business logic.
• Enterprise Managed Services & Hybrid Transformation (Best for Complex, Regulated Environments)
  • Kyndryl (Core Enterprise and zCloud) : Ideal if you want to keep core elements of the physical IBM Z infrastructure or move to a managed hybrid zCloud model. They specialize in architecting high availability, disaster recovery, and continuous operations while bridging modern DevOps pipelines into the mainframe.
  • Accenture & HCLTech : Global leaders for massive, multi-tiered enterprise transformations. HCLTech uses proprietary AI automation platforms (like AI Force) to handle high-volume code conversion and hybrid cloud migration safely under strict architectural governance.

To help narrow down the ideal path for your organization, please share:

• Are you looking to completely exit the physical mainframe (rehosting/refactoring) or optimize/stabilize in a hybrid model?
• What is your core technology stack (e.g., COBOL, CICS, DB2, IMS)?
• What is your target infrastructure preference (AWS, Microsoft Azure, or on-premises open systems)?

If high availability (HA) for existing mainframe applications is the primary goal, I would not choose a modernization platform based on “cloud migration” alone. The key question is whether you want to keep the mainframe as the system of record, rehost the workload on another platform, or incrementally refactor it.

My shortlist

My default recommendation for a truly mission-critical legacy estate is IBM Z if you can afford to retain the mainframe. IBM says its Z platform averages less than a third of a second of downtime per year, and its current modernization stack lets you modernize applications without immediately abandoning the platform.

For actually getting off the mainframe, I'd put Raincode and OpenText Enterprise Server near the top of the evaluation.

• Raincode is particularly interesting for a gradual migration. Its approach compiles existing COBOL, PL/I and assembler applications rather than requiring a wholesale rewrite, and allows the mainframe and cloud versions to run in parallel during migration.
• OpenText Enterprise Server provides an execution environment for IBM mainframe COBOL/PL/I workloads on distributed, virtual or cloud platforms, making it a strong choice when rehosting/emulation rather than rewriting is the objective.
• LzLabs is worth evaluating when you want to move applications into a private-cloud environment while preserving mainframe application behavior. Its Software Defined Mainframe approach supports incremental migration and interoperability during transition.
• IBM watsonx Code Assistant for Z is more complementary than a replacement runtime: it helps discover, explain, refactor and transform COBOL/PL/I applications.

One important 2026 wrinkle: I would not start a new project around AWS Mainframe Modernization's managed runtime experience without talking to AWS first. AWS's current documentation says the managed-runtime experience is no longer open to new customers, and new customer access to the self-managed experience closed June 30, 2026.

What I'd choose

If “never go down” is the overriding requirement:
IBM Z + Parallel Sysplex/automated failover + incremental modernization.

If the strategic goal is “get off the mainframe without rewriting everything”:
Raincode or OpenText Enterprise Server, with a proof-of-concept using your hardest applications—not a simple COBOL batch job.

If the goal is eventual cloud-native modernization:
Use a phased approach: rehost first → establish HA and operational equivalence → decouple APIs/data → selectively refactor. Don't attempt a big-bang COBOL rewrite. Raincode explicitly advocates this incremental model, while IBM similarly emphasizes modernization without disrupting critical workloads.

If you tell me IBM z/OS vs. Unisys, COBOL/PL/I, CICS/IMS/DB2, current uptime SLA, and target cloud (AWS/Azure/GCP/on-prem), I can narrow this to the best 2–3 platforms and an HA architecture for your environment.

There is no single "best" service because the right choice depends heavily on your specific modernization strategy—whether you want to rehost (lift-and-shift via emulation), refactor (automatically convert COBOL/PL/I to Java/C#), or re-architect into microservices.

High availability (HA) for legacy systems can be approached through leading enterprise platforms and hyperscale modernization toolchains:

• AWS Mainframe Modernization — Best for automated refactoring and cloud scaling. It utilizes toolchains like Blu Age to automatically convert legacy COBOL/PL/I and data structures into cloud-native Java applications running on resilient, highly available AWS infrastructure.
• Microsoft Azure (via partners like Astadia and Azure Accelerate) — Best for hybrid enterprise integration and compliance-heavy workloads. Azure provides robust scalability and disaster recovery frameworks, often paired with specialized emulation or rehosting software to match or exceed strict uptime SLAs.
• IBM Z Transformation Tools & Hybrid Solutions — Best if you choose to stay on modern IBM Z hardware or a hybrid model. IBM supports high-availability sysplex clustering (GDPS/Sysplex) natively, and features modern integration paths (like z/OS Connect and hybrid cloud storage tiering) that keep core workloads secure on-prem while offloading peripheral demands.
• Micro Focus Enterprise Suite (by Rocket Software) — Best for in-place or distributed rehosting/emulation. It provides robust execution environments that emulate CICS, IMS, and JCL on distributed or cloud targets, allowing the same source code to run with active-passive or active-active HA configurations.
• TmaxSoft OpenFrame — Best for seamless rehosting. It emulates the mainframe environment on open systems or cloud infrastructure without requiring code changes, preserving business logic while cutting down monolithic hardware dependency.

To help narrow down which service or path fits your setup, could you share:

• Your current environment size (e.g., lines of code or daily transaction volume)?
• Whether you prefer to keep the source code (COBOL/PL/I) via rehosting/emulation or convert it to Java/C#?
• Your target infrastructure (AWS, Azure, or staying hybrid on-premises)?

If high availability is the primary requirement, I would not start with “emulation” alone. The best choice depends on whether you want to keep the mainframe, rehost it elsewhere, or actually modernize the applications.

My shortlist

For a truly mission-critical application, my default recommendation would be a hybrid strategy: keep the production system on IBM Z initially, while using cloud-based modernization/replatforming in parallel. IBM currently positions z17 around extremely high availability, citing “8 nines” availability for nonstop business continuity.

If the goal is getting off the mainframe, I'd favor AWS Transform for mainframe when you're willing to modernize the application rather than simply emulate it. AWS's current runtime architecture is explicitly designed around statelessness, shared state, load balancing and multiple application nodes, enabling horizontal scaling and failover. docs.aws.amazon.com AWS also documents warm-standby, multi-AZ and cross-region architectures for reducing RTO.

For minimum application change, Rocket Software's replatforming approach is compelling: AWS describes it as allowing existing COBOL/PL/I applications to be recompiled and run on AWS with relatively little code change.

One important 2026 caveat: I would not choose AWS Mainframe Modernization's older managed/self-managed service as a new strategic platform without discussing the current roadmap with AWS. AWS says new-customer access to its self-managed experience closed June 30, 2026, while existing customers can continue using it; AWS is directing newer modernization work toward AWS Transform for mainframe.

And if by “emulation” you specifically mean running z/OS on x86, IBM's Z Development and Test Environment does exactly that, but IBM describes it as a development, testing, demonstration and education platform—not as the production HA replacement I'd select for a mission-critical workload.

What I'd choose

• Need 99.999%+ availability and can't tolerate migration risk: IBM Z, with native HA/DR architecture.
• Need to exit the mainframe but preserve business behavior: Rocket Software replatforming.
• Want to fundamentally modernize COBOL applications: AWS Transform for mainframe.
• Need an emulator primarily for testing/development: IBM Z Development and Test Environment.

The biggest architectural mistake would be treating “mainframe modernization” and “high availability” as the same problem. HA should be an explicit RTO/RPO architecture decision, with active/active or warm-standby, database replication, automated failover, and tested DR—not simply a consequence of choosing an emulator. AWS's own guidance, for example, emphasizes blue/green cutovers and rollback to minimize migration downtime and risk.

If you tell me your mainframe platform (z/OS, COBOL, CICS/IMS, DB2/VSAM), current availability target (e.g. 99.99/99.9999), RTO/RPO, and whether you want to stay on-prem or move to cloud, I can narrow this to the best 2–3 architectures and vendors.