HA/DR that is engineered, rehearsed, and operated — not just installed.
Replication design, switch testing, and operational ownership of IBM i and IBM Power estates that need to fail over in minutes, not headlines. Delivered by AS Networks engineers — built on Maxava, the platform we partner on for the most demanding IBM i environments.
Two disciplines, one resilience posture.
High availability keeps you running through component failure. Disaster recovery brings you back after a site-level event. Serious estates need both, layered — and we engineer them as a single posture.
High Availability
Keep production running through hardware, OS, and application failure. Local or metro replication, sub-second data loss, role swap in minutes. The instrument used to take planned maintenance windows off the calendar.
- +Local / metro replication
- +Sub-second RPO for journalled data
- +Role swap measured in minutes
- +Planned-maintenance switch as a non-event
Disaster Recovery
Resume operations after a site loss, prolonged outage, or ransomware event. Geographic replication, defined RPO/RTO, rehearsed runbook. The instrument that turns a board-level event into a tested procedure.
- +Cross-region replication
- +Bounded RPO under contract
- +Recovery from rehearsed runbook
- +Auditor-ready evidence each quarter
The whole IBM i object surface — not just the database.
A swap that brings the data over but leaves the user profiles, authorities, or output queues behind is a swap that fails the next morning. We engineer the full object scope from day one.
DB2 for i
Physical and logical files, journals, and journal receivers. Real-time replication via the IBM i remote journaling stream — the platform's most efficient transport.
Integrated File System
Multi-threaded IFS replication for document stores, web roots, and Java workloads — scaled up to hundreds of parallel streams on high-volume estates.
Libraries & objects
Programs, service programs, data areas, data queues, job descriptions, output queues, and the supporting objects an application actually needs to start.
Security objects
User profiles, authorities, authorization lists, and system values — replicated on cadence so the target can actually be used the moment it's swapped to.
Spool & work management
Spool files, job queues, subsystem definitions, and scheduling — so the operational picture on the target matches the source, not a blank installation.
Adjacent platforms
AIX and Linux on Power covered for monitoring, runbook orchestration, and coordinated failover — so the dependent stack moves with IBM i, not after it.
Journal-level replication, continuously audited.
Customers buy confidence in the mechanism, not the marketing. Here is the engineering we stand up — built on Maxava and IBM i remote journaling, the most efficient transport the platform provides.
Remote journaling
The IBM i journal stream is shipped to the target system in real time. Native to the platform, OS-level, and designed for transaction-heavy workloads. Sub-second lag in steady state.
Concurrent apply, scaled to workload
Apply processes on the target replay journal entries in parallel — sized to your transaction profile so backlogs don't accumulate during peak windows or batch runs.
Before / after audit
Continuous record-level comparison between source and target. Drift is detected as it happens, not at the moment a swap is required. Issues raised as alerts, not as failed failovers.
Object-level cadence
Programs, system values, and user profiles — the objects that don't journal — are replicated on a defined cadence so the target stays a usable LPAR, not just a data dump.
Three variants. One rehearsed runbook.
The same scripted procedure runs whether the swap is planned, simulated, or unplanned. What you test is what executes.
Planned
Operator-initiated for hardware refresh, OS upgrades, or firmware work. Production runs on the target while the source is touched. Reverse when done — no extended outage required.
Simulated (SRS)
Maxava Simulated Role Swap proves the DR plan works without taking production down. Run quarterly as standard — and documented to a standard your auditor can sign off on.
Unplanned
Real incident. Scripted commands handle IP, DNS, application config, and user redirection. The procedure is the same one that ran in last quarter's SRS — no improvisation at 3am.
Engineered outcomes, not vendor promises.
Final RPO/RTO targets come out of the discovery and design phase against your actual workload. As a baseline, here is what a well-engineered IBM i HA/DR practice delivers.
< 1 second
Sub-second data loss tolerance in steady state for journalled data. IFS-heavy workloads target under five seconds depending on change rate.
5 – 15 minutes
Time to be running on the target for a well-rehearsed swap. Larger estates and those gated by external systems (DNS, identity, third-party) sized during design.
Quarterly SRS
Simulated Role Swap as standard cadence — monthly for regulated customers. Documented evidence package included with every run.
Three engagement models. Self-select on operating burden.
Software, customer-operated
We license, install, and configure Maxava on your hardware and hand over to your team for day-to-day operations. AS Networks backstops with quarterly health checks, SRS support, and senior escalation. Lowest cost; highest customer ownership.
Co-managed
Our NOC shares day-to-day monitoring with your team. We handle remediation, journal hygiene, capacity planning, and SRS execution; you retain decision rights on swap and operate the production application. The default model for most mid-market estates.
Fully managed — DR as a service
AS Networks provides the DR target capacity, operates replication, monitors 24/7, runs SRS quarterly, and executes the swap on your authorisation. The right model when IBM i operations skill no longer sits in-house.
What a service practice does that a software vendor can't.
Maxava is the platform we run. AS Networks is what operates around it — and operates the parts the software cannot reach.
Runbook authorship
The software ships the swap engine. We write the runbook for your specific stack — including the third-party systems that always get forgotten until 3am: DNS, certificates, MQ queues, payment gateways, AD service accounts.
Application-aware testing
SRS proves the OS swaps. We prove the application still works after the swap — order entry, scheduled batch, integration flows, customer-facing endpoints. A green light from us means the business works, not just the platform.
Failback discipline
Most disasters are followed by a botched failback. We treat failback as a first-class operation with its own scripted runbook, rehearsed alongside the forward swap.
Compliance evidence
Quarterly SRS reports, drift logs, lag histograms, RPO/RTO measurements — packaged for ISO 27001, SOC 2, PCI-DSS, and financial-services auditors. Your CISO stops chasing screenshots from the operations team.
Capacity planning
Journal receivers, apply backlogs, IFS growth, target-side storage. Tracked as trends, raised as advisories — not discovered the day the system can no longer keep up.
Maxava partner, vendor-experienced
We are a Maxava partner and run it as our reference platform. Our engineers have also operated MIMIX, iTera, and native IBM PowerHA estates — so if you are inheriting one, we can operate it before deciding whether to change it.
What an untested HA/DR posture costs when it's tested for you.
Unplanned downtime cost
Mid-market IBM i estates typically lose £5k–£50k per hour of unplanned outage. Transaction-intensive sectors — retail at peak, payments, logistics — pass £100k/hour.
Recovery time without rehearsal
Backup-only postures typically take 6–24 hours to restore production-grade service. A rehearsed HA/DR practice delivers the same outcome in 5–30 minutes.
Compliance exposure
Financial-services, healthcare, and government auditors increasingly require tested DR — not documented DR. Untested runbooks are findings, not controls.
Ransomware blast radius
A synchronously replicated HA pair without journal isolation and offline copies is a single blast radius. Real HA/DR posture isolates the recovery path from the production path.
Common engagement shapes.
DR readiness assessment
A short engagement that inventories your current posture, measures the gap to a tested HA/DR baseline, and lands a sized proposal — not a slideware promise. Typically two to four weeks.
Ready to model what independent support looks like for your estate?
Fixed-fee two-week assessment. We deliver a renewal-trajectory model, risk grading, and a written recommendation — yours to keep, no obligation.