Based on the AWS Storage Blog case study (12 October 2026). Performance and cost figures are the vendor- and customer-reported results in that post — treat them as claims to validate, not independent benchmarks.
FSX FOR ONTAP PERFORMANCE ENERGY 2026-10-12Halliburton Landmark moved the subsurface storage behind its iEnergy® cloud platform and Landmark™ Geosciences Suite off a self-managed NAS onto Amazon FSx for NetApp ONTAP. The AWS Storage Blog case study published on 12 October 2026 reports roughly 45% overall performance gain, ~5x faster 3D seismic data loads, ~10x faster interactive 3D visualisation, and 15–20% lower storage cost than the previous solution. NetApp holds AWS Energy Competency for this segment.
Why the switch: seismic volumes outgrew the old NAS
Landmark delivers Geosciences Suite, a DecisionSpace® 365 solution, as a cloud service on AWS. Its customers work with seismic data plus the associated interpretations, wells and reservoir data — and those datasets keep growing. The team cited rising seismic data volumes, a need for near-real-time processing of large volumes, low-latency access to historical data for machine-learning workflows, and the capital, maintenance and specialist-staff cost of running storage themselves.
FSx for ONTAP was selected as a managed, multi-protocol file system that serves both NFS (Linux) and SMB (Windows) clients, with migration options and ONTAP's data-management features built in.
The reported performance numbers
- ~45% overall performance gain across operations.
- 3D seismic data load times reduced by about five times.
- Sustained tens of GB/s of throughput, with performance holding up when multiple geoscientists work concurrently.
- Millions of IOPS addressing a previous bottleneck in daily workflows.
- Visualisation, UI response and interactive 3D work such as seismic scrolling, panning and stepping about ten times faster; smoother QC and faster horizon/fault picking.
The I/O pattern behind the numbers
The reason a general-purpose NAS struggled is the mixed read profile of subsurface interpretation. The post breaks the workflow into sub-workflows, each with its own demand on the file system:
- Seismic data loading and QC — large sequential reads with heavy I/O bursts while formatting and validating.
- Horizon and fault picking — frequent random reads of small-to-medium blocks as a geoscientist navigates inlines, crosslines and time slices.
- Attribute computation — high-throughput sequential I/O that spawns multiple derived sub-volumes.
- Seismic visualisation — mixed random and sequential reads at high concurrency.
- Well-seismic ties — moderate random access over small windows around wells.
- Reservoir modelling — sequential export to simulation, with random reads for updates and calibration.
That workload shape is exactly what ONTAP admins plan capacity and tiering around — a mixed random/sequential profile that rewards the right SSD/capacity balance and steady-state monitoring (see performance monitoring).
Operations and cost
- A targeted ~30% reduction in server maintenance time, as patching and updates move into the managed service.
- About 15–20% reduction in storage cost versus the previous storage solution.
- Intelligent automatic data tiering between high-performance SSD and lower-cost capacity storage, driven by access patterns.
- Compression and deduplication to shrink the storage footprint.
The optimisations run in the background, so iEnergy customers do not see workflow disruption.
Backup, disaster recovery and resilience
- NetApp SnapMirror used for cross-Availability-Zone replication, creating space-efficient copies across AZs.
- Asynchronous replication for continuous protection without hitting primary performance.
- A target recovery time objective (RTO) of under 10 minutes.
- A target disaster-recovery storage cost reduction of up to 50%.
These are the same primitives ONTAP operators already use on-premises: SnapMirror replication and a multi-AZ footprint are a familiar route to a resilient disaster-recovery design.
What it means for ONTAP operators
The headline here is not that cloud storage is fast — it is that the capabilities doing the work (SnapMirror, tiering, deduplication, compression, multi-protocol NFS/SMB) are the same ONTAP features an on-prem admin manages daily. The FSx for ONTAP deployment is a template: identify the I/O profile per workload, tier accordingly, and lean on replication rather than a bespoke DR build.
Two caveats worth holding onto. First, every ratio above is customer- and vendor-reported on their own datasets, not an independent benchmark — seismic throughput depends heavily on file sizes, client count and network path. Second, the "cost reduction" figures compare against Landmark's previous self-managed environment, so they do not translate directly to a managed-versus-managed comparison. The useful takeaway is the architecture, not the percentages.
“Our focus is to advance the performance, reliability, and functionality of our seismic interpretation solutions. Amazon FSx for NetApp ONTAP supports that objective by improving application performance and allowing our engineering teams to dedicate more time to innovation and customer-focused development.”
— Robert Kellogg, Director of Innovation and Marketing, Halliburton Landmark
Sources: AWS Storage Blog — “How Halliburton Landmark enhanced seismic interpretation performance and data management using Amazon FSx for NetApp ONTAP” (12 October 2026) · Amazon FSx for NetApp ONTAP