Expert Guide IBM Storage Protect: RESA Client, Defender Integration, and the Strategy Following the GSE ISP Symposium 2026— Review

Andreas Wagener — 17. Nov 2026
Reading time: 5:53 minutes

Die diesjährigen Symposium-Themen unterstreichen, dass Cyber-Resilienz inzwischen technische, organisatorische und regulatorische Ebenen gleichzeitig betrifft.

Reading from the GSE ISP Symposium 2026: How the new Resiliency Agent (RESA), the connection to IBM Storage Defender and DORA compliance shape the future of IBM Storage Protect - a classification for admins and decision-makers

“Is Storage Protect still a thing?” – IBM representatives were confronted with this question in customer discussions at the 18th GSE ISP Symposium in Leipzig. The answer followed in the form of solid technical architectural decisions: A newly developed client called Resiliency Agent (RESA) as well as the deep connection to the higher-level cyber resilience platform IBM Storage Defender (we described the last update here) show the targeted modernization, and not a possible phase-out of the backup platform.

Targeted modernization of IBM Storage Protect

The symposium covered a total of five major topics:

  • The integration with IBM Storage Defender
  • Tape as a reliable building block of cyber resilience
  • The roadmap around the new RESA client
  • Changes to support and security policies
  • The practical use of AI in the backup environment

This review focuses on the first two developments - RESA and integration with IBM Storage Defender - but also groups the remaining topics as together they provide a more complete picture of the current development of IBM Storage Protect.

Where the uncertainty surrounding IBM Storage Protect comes from

We were hardly surprised by the skepticism of IBM customers: The hesitant further development of IBM Storage Protect has been noticeable in our projects for years - less because customers are sticking to old versions, but because for a long time there was hardly any dynamic in the product itself. In its own session in Leipzig, IBM structured its response along five feedback points from the last symposium:

  • Support
  • Up-to-dateness of supported platforms
  • Functional extensions
  • Release rhythm
  • Strategic focus

IBM pointed to a reduced ticket backlog, more than 70 newly supported operating system and application versions within 18 months and four releases within a year as progress.

New currency policy and release strategy for clients

A new currency guideline for clients and so-called TDPs (data protection modules for applications such as databases) are also under discussion. This is based on the support life cycle of the respective manufacturer and will in future differentiate precisely between actively supported, expiring, withdrawn and no longer supported versions. A comparable guideline for the server itself will follow at a later date.

In addition, an annual “launchpad release” with its own license file is planned – initially only for clients, not for the server. There was clear criticism from the audience: an annual mandatory upgrade is hardly realistic with multiple servers with dependent applications and narrow maintenance windows. The desire for real long-term support versions (LTS) was expressed several times, especially for the server area.

Concrete pressure to act in practice

The fact that this pressure to act is by no means just theoretical is demonstrated by concrete technical framework conditions:

  • Db2 11.5 end of support: Standard support for the Db2 base of IBM Storage Protect 8.1 ends on April 30, 2027, according to the official IBM Support Lifecycle page, while Db2 12 cannot be easily backported to version 8.1.
  • GSKit 9 & Strict TLS 1.3 mode: The switch to GSKit 9 and strict TLS 1.3 operation can only be implemented in stages for large client fleets - calculated using the example of 4,500 clients in the lecture. For very old version 5 or 6 clients without SSL support, the connection still works today as long as the server offers a classic non-SSL port in addition to the SSL port and server-side options such as SSLDISABLELEGACYTLS are not set to strictly enforce TLS 1.2 or higher. With the announced strict TLS 1.3 mode, this leeway is completely eliminated, so that outdated clients are then consistently locked out.

The new client: Resiliency Agent (RESA) – The update for admins

At the center of the announcements is a completely newly developed client for IBM Storage Protect, internally referred to as Resiliency Agent (RESA). The aim is to replace the historically grown clients, which are structured differently depending on the workload, where functions such as instant mount or a ransomware check previously had to be developed separately for each workload.

The four design principles of RESA

  • Unified format: A single data format for all workload types.
  • Direct mounting: Each backup can be mounted directly via a file system mechanism (FUSE layer) without data having to be copied back in advance.
  • Snapshot technology: Each backup is technically processed as a snapshot.
  • Immutability: Every snapshot created is natively immutable (unchangeable).

Workloads, platforms and architecture

RESA initially covers three primary workload areas:

  • File systems as a standard case (including automatic, intelligent splitting of very large Ceph file systems).
  • Object storage across defined target buckets.
  • The PostgreSQL and MongoDB databases.

The initial platforms planned are Linux (x86 and Power) and AIX; Versions for Windows and macOS are in internal development.

Technically, every backup at RESA goes through a uniform processing chain: deduplication, threat and anomaly detection, compression and encryption. Each agent creates its own local repository and works with cumulative snapshots. A “clone-and-mount” function already allows backups to be read-mounted; writable clones are still in development. The interface is entirely REST API based - the command line is just a thin wrapper over it.

Integration into existing environments & licensing

Important for existing environments: RESA continues to use the existing IBM Storage Protect servers (IBM Storage Protect 8.2 is named as the target server), their policies and schedules. No new storage pool types are required. When it comes to licensing, a PBU model is initially provided for on-premises databases (licensing per processor core or unit), and a terabyte-based license is used for the remaining workloads. According to the spokesman, more detailed information will follow within the next two quarters. The current year was given as the time frame for initial availability – as a technology preview.

The question round resulted in open points for practice: The consistency mechanism for snapshots is still workload-dependent (for Windows, for example, a connection to VSS is planned). The actual backup is currently file-based and not as a block image. How exactly the snapshot time is determined when there is no native source snapshot remains to be clarified.

Redevelopment of the operations center

In parallel with RESA, IBM is also announcing a completely new development of the IBM Storage Protect Operations Center. The new portal is intended to bundle operations, administration, service and monitoring in a common interface. In the long term, the aim is to centrally manage multiple RESA services via the same REST API base in the sense of a “single pane of glass”.

Integration into IBM Storage Defender: Cyber Resilience & DORA Compliance (Decision Maker Perspective)

The symposium also clearly positioned IBM Storage Protect within IBM Storage Defender, the overarching cyber resilience platform. For managers, the focus is no longer purely on data backup, but rather on the verifiable recoverability of digital business processes under regulatory pressure. IBM Storage Protect therefore increasingly acts as an execution layer beneath IBM Storage Defender.

IBM describes the interaction along a four-role model - C-level, application managers, storage admins and backup admins - who require different information in an emergency but now often work separately.

At a glance

The Data Resiliency Service (DRS) brings together the perspectives of C-level, application and storage managers:

  • Application Groups: Mapping technical dependencies for the correct order of recovery of complex systems.
  • Anomaly detection: FlashCore Modules (FCM4) and Defender sensors detect threats in real time and automatically trigger untamperable safeguarded copies in the event of anomalies.
  • Data Management Service (DMS): The cloud portal enables multi-client reporting and secures critical administrative actions using a strict four-eyes principle.

Application Groups

Technical dependencies are mapped via so-called recovery groups and application groups - which systems belong together and in which order they need to be restored. Governance profiles check specifications such as the 3-2-1 rule across manufacturers. Restores can be tested automatically in an isolated environment (clean room), instead of just once a year as part of an exercise. According to the spokesperson, the service is updated monthly; a 60-day trial is available free of charge.

Anomaly detection

A sensor from IBM Storage Defender and the IBM FlashCore modules (from the fourth generation) work together for early detection: If the sensor detects an anomaly, a so-called safeguarded copy is automatically created on the storage system. It was expressly emphasized that DRS does not stop an ongoing attack itself, but rather aims for a quick, verified recovery.

Data Management Service (DMS)

IBM Storage Defender Data Management Service (DMS) runs in the cloud and collects inventory, reports and policies from multiple Data Protect clusters in one place. Sensitive actions can be secured with a four-eyes principle: a second person must agree before they are carried out. Several customers or areas can be clearly separated from each other; there are so-called “organizations” for this purpose. Technically, outgoing HTTPS traffic is sufficient for operation. If you also want to scan backups for threats using YARA rules, you must license this function separately.

Other technologies: Tape as an air gap, automation & AI

At a glance: Tape as a physical air gap

  • S3 interface (IBM Deep Archive): Enables S3 protocol access to tape libraries (such as the Diamondback with up to 60 PB). A recall only takes 1 to 3 minutes (compared to up to 48 hours for cloud archives on AWS).
  • Safeguarded Tape: Creates a logical partition invisible to backup applications. Exported cartridges are automatically isolated and protected from delete commands.
  • LTO-10: New read/write heads eliminate the previous cassette initialization and increase temperature and moisture resistance.

S3 interface (IBM Deep Archive)

In this session, IBM provided a technology and roadmap update on two existing systems. The Diamondback Library is a 19-inch rack library with up to 15,000 slots, up to 14 drives and a capacity of up to 60 petabytes.

IBM Deep Archives

IBM Deep Archive is technically based on this: a turnkey, S3-compatible tape-based interface that is primarily aimed at environments without a native tape connection. Access is via the S3 protocol instead of Fiber Channel. Before the actual data access, an upstream recall step is necessary – similar to AWS Glacier. This takes around one to three minutes, while AWS Glacier requires a recall of up to 48 hours.

Safeguarded tape

With cyber resilience in mind, the feature creates a logical partition, invisible to backup applications, into which exported cartridges are automatically moved so that they are inaccessible to delete commands from within the application. It runs automatically and can be used free of charge, but requires firmware from October 2025. For additional security, a locally protected mode can be activated, in which only an administrator can put cartridges back into a production partition via the local GUI on the library.

LTO-10

On the drive side, LTO-10 brings two technical improvements: An inclined, rotating head adapts to the tracks, which eliminates the previously necessary initialization of the cassettes. A more corrosion-resistant read/write head should allow operation even at higher temperatures and humidity - albeit with a staggered price increase that IBM introduced for tape and hard drive products over the course of the year. For forensic evaluation of tape operation, 350 to 400 data points can be recorded per mount process and evaluated in Grafana, for example.

A case from the Leibniz Computing Center (LRZ) showed how useful this is in practice.

Monitoring there revealed dust in a tape library as well as a faulty LTO-9 batch with a tape thickness outside of tolerance.

Two points remain open: Firstly, tape reclamation - i.e. merging still valid data from several partially occupied cassettes onto new tapes in order to reclaim storage space - is still running too slowly in his opinion. On the other hand, restoring from locally protected container storage pools has not yet been solved satisfactorily, which is particularly relevant for environments with container-based backup targets that also store their data on tape.

Compliance classification – DORA (Digital Operational Resilience Act)

A dedicated IBM Solutions team systematically tested and classified IBM Storage Protect against nine articles of the EU DORA regulation.

DORA Article 

Coverage Level 

Technical Implementation in IBM Storage Protect 

Article 7 (Resilience) 

Directly covered 

High-availability clustering & automatic failover. 

Article 9 (Protection & Prevention) 

Directly covered 

AES-256 (at rest), TLS encryption, RBAC, WORM / S3 Object Lock. 

Article 10 (Detection) 

Directly covered 

End-to-end audit logging and SIEM integration (e.g. IBM QRadar, Splunk). 

Article 11 (Business Continuity) 

Directly covered 

Multi-site replication, RTO/RPO policies, point-in-time recovery. 

Article 12 (Backup & Restore) 

Strongest coverage 

3-2-1-1 copy rule, geographic redundancy, automated test restores. 

Articles 6 & 8 (Risk Framework/Assets) 

Partially covered 

Requires integration with external GRC and asset management tools. 

Article 17 (Incident Management) 

Partially covered 

Provides raw data, but requires an ITSM/SOAR platform for regulatory reporting. 

Article 21 (Resilience Testing) 

Partially covered 

Automated restores available; penetration tests (TLPT) are not covered by the product. 

An official document on the DORA analysis has been announced.

There was controversy in the discussions as to whether a central LDAP/AD connection for backup systems is advisable. While US compliance regulations require central user management, many European admins prefer isolated local emergency accounts so that an AD compromise does not block access to recovery points.

Storage Protect in a Box – automated long-term archiving

With "Storage Protect in a Box", IBM shows how IBM Storage Protect can be used as an archive for IBM Storage Defender and Copy Data Management: Data that was initially only backed up for a short period of time can be stored for the long term. It is not a hardware appliance, but a collection of Ansible playbooks that reduce installation and configuration to a few commands and build a complete instance including hardening within 45 minutes.

AI – opportunity and risk

Using a Model Context Protocol (MCP) server, IBM Storage Protect can be controlled using natural language using AI models (such as Claude or ChatGPT Enterprise). Functions such as server monitoring or searching for backup statuses before a ransomware infection can be triggered in this way. Writing actions require approval via the SkillGuard security concept, isolated sandbox execution and a four-eyes principle. Integration into the operations center is planned for mid-2027. AI in backup operations is still met with reluctance by users - especially with regard to prompt injection, data sovereignty and compliance rules.

IBM Storage Protect in practice – use case from the Leibniz data center

Two lectures from the Leibniz Computing Center showed how IBM Storage Protect actually operates in very large environments.

  1. Database backup: Concept for backing up IBM Storage Protect's internal database across many server instances, using a mix and tape as an air gap.
  2. Storage Scale Migration: Migrating the internal database and log files to IBM Storage Scale with fileset-based structures to make snapshots more usable and bring physical server instances closer to managing virtual instances.

Classification

The symposium showed that the IBM Storage Protect product is being systematically modernized and repositioned as a core component of modern cyber resilience - and that IBM is clearly listening to the various user groups. In the short term, the main thing to check is your own server version, as the Db2 version on which 8.1 is based will soon be out of support.

In this context, the new release concept for clients and how well it can actually be implemented by users remains exciting - this will have to be monitored over the next year. The introduction of the RESA client is in any case a strong signal for the support of additional workloads.

The presentations at the symposium also showed how IBM Storage Protect is used in large environments and how the product must be integrated into increasingly complex infrastructures in order to meet companies' growing security needs.

Other topics at the symposium - tape, compliance and AI - made it clear that the current change not only affects the client, but also affects the infrastructure level, regulatory proof requirements and the operating model.

Conclusion: People who are said to be dead live longer

During the discussions in and around the program contributions, there was clear interest from both IBM and IBM end customers and business partners. The new RESA client for new workloads and a new operations center provide a strong signal, while connectivity to IBM Storage Defender and automated CleanRoom testing pave the way for modern DORA and compliance requirements. This year's symposium topics underline that cyber resilience now affects technical, organizational and regulatory levels simultaneously. Whether and when this has a specific impact on your own environment depends heavily on the current version, the workloads used and the ability to upgrade in your own company.

You were interested in this, then you may also be interested in...