Author: Rob Pankow
Showing 120 posts.
Rob Pankow
Nov 11th, 2026 | 7 min read
Edge Modernization for Stores and Warehouses: Designing the Two-Server Site
An edge fleet is one small site design repeated many times. Decide what must stay on site, whether a site gets one server or two, and how two servers stay available without a storage witness at the site, before you pick the pilot sites.
Rob Pankow
Nov 11th, 2026 | 6 min read
A Complete Storage Package for OpenShift Virtualization: What to Ask For
VM disks, recovery, shared files, and object storage are often bought as four products. On OpenShift they can come from one storage platform, and the limits of that answer are worth knowing before you ask for it.
Rob Pankow
Nov 11th, 2026 | 7 min read
Your Storage Works in the Lab. What to Test Before VMs Run in Production
A lab shows that storage installs and a VM boots. Before production, test a node drain, a node failure with its rebuild, performance under real load, and a full recovery, each against a limit you set in advance.
Rob Pankow
Nov 11th, 2026 | 8 min read
Moving from VMware to OpenShift? You Can Keep Your SAN on Day One
The array under your vSphere cluster can stay when the VMs move to OpenShift. A storage layer on the OpenShift hosts runs on new LUNs from the same array, so you migrate the VMs now and decide about the hardware later.
Rob Pankow
Nov 7th, 2026 | 9 min read
vSphere Replication RPO After the VMware Exit: Where the Setting Lives Now
Granularity moves from the VM to the volume, and the interval is a cadence rather than a ceiling. A StorageClass per RPO tier carries the setting, and you check lag yourself.
Rob Pankow
Nov 4th, 2026 | 6 min read
vCenter vs vSphere: What Each One Is and What Replaces It
ESXi is the hypervisor, vCenter Server is the management server, and vSphere is the suite around both. After the exit, hosts become worker nodes and vCenter's jobs spread across several tools.
Rob Pankow
Oct 31st, 2026 | 8 min read
VVF vs VCF: What Storage vSphere Foundation and Cloud Foundation Include
vSphere Foundation and Cloud Foundation ship the same core vSAN features. VCF differs in a cyber recovery add-on VVF cannot take, principal storage rules for its domains and, under VCF 9.0 terms, a larger included vSAN capacity.
Rob Pankow
Oct 28th, 2026 | 10 min read
Thin vs Thick Provisioning: What Each VMDK Format Becomes on Kubernetes
A VMDK's provisioning format does not carry over to Kubernetes. Allocation follows the storage, CDI preallocation is opt-in, and quotas and storage caps stand in for the space a thick disk reserved.
Rob Pankow
Oct 25th, 2026 | 11 min read
Rack Failure Domains and Erasure Coding: How Many Racks Each Scheme Needs
Your rack count decides which erasure coding schemes keep running through a full rack outage. simplyblock sets failure domains when the cluster is created, so pick racks, scheme and hosts together before you order hardware.
Rob Pankow
Oct 22nd, 2026 | 10 min read
How to Expand a VM Disk on OpenShift Virtualization While the VM Runs
Raise the size of the PVC behind the disk, and the CSI driver and KubeVirt grow the volume under the running VM. Then extend the partition and file system from inside the guest.
Rob Pankow
Oct 14th, 2026 | 13 min read
Disaster Recovery Testing After the vSphere Exit: How to Rehearse a Failover You No Longer Have a Button For
A disaster recovery plan that has never been executed is a hypothesis. On vSphere, rehearsing it was one button, because the array handed back a writable snapshot of the replicated device and mounted it to an isolated port group. After the exit that guarantee is assembled from primitives instead: a snapshot of the replicated volume, a clone taken from it, an isolated namespace, and a teardown. Here is the drill in order, what each step has to prove before the next one is meaningful, and how to roll it back.
Rob Pankow
Oct 5th, 2026 | 14 min read
vSAN ESA vs OSA: What the Express Storage Architecture Actually Requires From Your Hardware
The honest difference between the Express Storage Architecture and the Original Storage Architecture is not a feature list. OSA was a software layer you could lay on hardware you mostly already owned. ESA is a hardware specification with software attached: all-NVMe TLC, no cache tier, a faster fabric, and a prescriptive node profile. That makes the move a procurement event, and it puts a quote for all-NVMe hardware in your hand. This post compares the two architectures on the five criteria that decide what you have to buy, then asks the question the quote actually raises.
Rob Pankow
Oct 1st, 2026 | 11 min read
A GPU-Attached VM Cannot Live Migrate: GPU Passthrough and vGPU After the vSphere Exit
The GPU boxes are usually the last wave in a migration plan, and they are the wave where the plan quietly changes shape. A virtual machine holding a physical GPU is not a normal VM with an extra device. The assignment itself is what removes live migration, and no storage layer gives it back. What storage does decide is how long every unavoidable stop lasts.
Rob Pankow
Sep 28th, 2026 | 15 min read
The Key Provider Does Not Travel: VMware Encryption Key Management After the vSphere Exit
Run the inventory and the encrypted machines are a subset, not the estate. That subset has a different migration path from everything around it, and the reason is not the cryptography. Both vSphere encryption features terminate in a Key Provider, and the Key Provider is the one component of the trust chain that has nowhere to land on the other side. Here are the three paths out, and what each one actually costs.
Rob Pankow
Sep 25th, 2026 | 13 min read
The VM Template After the vSphere Exit: Three Ways to Run an Image Catalog
A VM template was never one thing. It was a catalog, a portable envelope and an identity step, sitting on top of a storage trick that made a new machine appear without copying the image. After the exit those jobs split apart, and there are three defensible places to put them. Which one is right depends on image size, how often you instantiate, and how many clusters need the same picture.
Rob Pankow
Sep 22nd, 2026 | 14 min read
Migrating VMware to Proxmox VE: The Storage Decision to Make First
Most guides to migrating VMware to Proxmox VE are compute guides. They cover the cluster build, the network bridges and the disk import, and they treat storage as a prerequisite bullet. It is not a prerequisite, it is the decision: the storage backend you pick determines whether live migration survives, whether you still have snapshots, and how much raw disk you buy to replace a licence bill. Here are the paths, what each one costs, and the condition that makes each one correct.
Rob Pankow
Sep 19th, 2026 | 13 min read
pNFS and Parallel RWX: The Shared Access KubeVirt Live Migration Actually Needs
ReadWriteMany is not one mechanism. It is two, and which one you get depends on the volume mode. For raw block devices the consumer arbitrates. For a shared filesystem something below has to supply namespace consistency, lock arbitration and cache coherence, and the usual answer funnels every client through one server head. This is how parallel NFS splits the control path from the data path, and why that split decides whether a KubeVirt VM can be live migrated at all.
Rob Pankow
Sep 16th, 2026 | 14 min read
The vTPM Becomes a PVC: Where Windows VM Firmware State Lives After the vSphere Exit
A vTPM is not a stateless device. It holds keys, and on vSphere that state lived inside an encrypted file in the VM's home directory, where nobody had to think about it. On KubeVirt the same state is a PersistentVolumeClaim with a hard volume mode requirement, an access mode requirement, and consequences when it is lost. This is what that volume actually is, what it needs from the storage layer, and the part of the migration no storage vendor can fix for you.
Rob Pankow
Sep 11th, 2026 | 16 min read
Leaving vSphere Without a Storage Refresh: What Happens to the Array You Already Own
Every exit plan starts with the hypervisor. Almost none of them start with the array sitting under it, which has a support contract, a depreciation schedule, and a refresh date of its own. Those two clocks are now running at the same time, and the decision about the second one is usually made by default. Here is what keeping the array actually buys, what it keeps costing, and how to tell which answer is yours.
Rob Pankow
Sep 8th, 2026 | 14 min read
Leaving VMware Tanzu: The vSphere CSI Driver Does Not Come With You
If you were already running Kubernetes on vSphere, your exit looks easier than everyone else's. Your PVCs are Kubernetes objects, your manifests are already declarative, and no workload needs converting. Then you look underneath and find the provisioner, and the provisioner is not portable. This is what actually happens to the storage layer when a Tanzu estate leaves, which pieces survive the move, and which ones only look like they do.
Rob Pankow
Sep 4th, 2026 | 13 min read
Storage Tiering After the vSAN Exit: What Decides Which Data Sits on Flash
vSAN made the hot-versus-cold decision for you, at the disk group, with a ratio you fixed at build time. Leaving takes that decision away and hands nobody a replacement. Storage policies do not fill the gap, because they declare capability rather than placement. Here is who actually decides what lives on flash afterwards, and why the return trip out of cold storage is the part that turns into an incident.
Rob Pankow
Sep 1st, 2026 | 13 min read
Data Residency Requirements: The Sovereignty Decisions Your Storage Layer Makes
Sovereignty appears in the exit business case as a slide, then disappears into the architecture. It re-emerges as six storage decisions that nobody planned: where replicas land, where snapshots go, who holds the keys, who can reach the data plane, where telemetry is exported, and where the DR site sits. Most of them are made by a default rather than by a person. Here is the list, what vSphere decided for you before, and which defaults can cross a boundary without anyone noticing.
Rob Pankow
Aug 29th, 2026 | 11 min read
The Other VMware Exit: Storage for Providers Who Move Customers, Not Workloads
Almost everything written about leaving VMware assumes you own the workloads. Hosting providers, MSPs and sovereign cloud operators do not. Their tenants are customers, their tenant boundary is a contract, and their migration happens one account at a time on schedules they do not control. That changes the storage decision completely.
Rob Pankow
Aug 26th, 2026 | 14 min read
What Happens to Your VMware Backups When You Leave vSphere
Every VMware exit plan budgets for moving the virtual machines. Almost none of them budget for moving the backups, and the backup estate is a second migration hiding inside the first one. Here is what the hypervisor backup API, changed block tracking, the proxy topology and the retention chain each become after the exit, which parts a storage choice decides, and which parts stay firmly the backup product's job.
Rob Pankow
Aug 23rd, 2026 | 15 min read
What Replaces vMotion and Storage vMotion When You Leave vSphere
Every VMware exit guide argues about where the VMs should land. None of them tell you what happens to the features your maintenance windows depend on. vMotion has a direct equivalent whose precondition is storage. Storage vMotion has none, and on a pooled storage layer most of what it was invented to solve stops existing. Here is the feature-by-feature translation, including the two that are not storage problems at all.
Rob Pankow
Aug 19th, 2026 | 13 min read
VMware Perpetual License: What Actually Expires, and What It Costs to Keep Running
The license is perpetual. The support contract is not. Understanding which of the two actually expires changes the renewal conversation, and it moves the real deadline from the vendor's calendar to your hardware refresh.
Rob Pankow
Aug 14th, 2026 | 10 min read
Cloud Repatriation: Why Data Gravity Decides Whether You Can Actually Leave
Compute is portable and everyone knows it. The reason repatriation projects stall is that the persistent data is not, and the cost of moving it is decided years earlier by the storage layer you chose.
Rob Pankow
Aug 6th, 2026 | 8 min read
OpenShift Virtualization Storage: The Kubernetes-Native Alternative to Commercial HCI Platforms
Commercial Kubernetes-native HCI platforms solve real problems for OpenShift Virtualization, but their separate, capacity-tiered storage subscription and proprietary control plane bring their own lock-in. What to check before committing to one.
Rob Pankow
Aug 5th, 2026 | 8 min read
The Double-Replication Tax: Rethinking Storage Redundancy for Databases After the VMware Exit
vSAN let teams dial per-VM storage redundancy with an explicit policy. Move a self-replicating database onto a Kubernetes storage layer with its own fixed replication factor, and that dial disappears, silently stacking a second redundancy layer on top of one the database already provides.
Rob Pankow
Aug 4th, 2026 | 16 min read
Best VMware Alternatives 2026
The best VMware alternatives in 2026 span three families: Kubernetes-native platforms, hypervisor-first platforms, and packaged HCI. Compare OpenShift, Rancher, Talos, Cozystack, Proxmox VE, Hyper-V, Nutanix, Harvester, and OpenStack, and see why the storage layer decides whether the migration holds up.
Rob Pankow
Aug 4th, 2026 | 6 min read
Platform Engineering After the VMware Exit: Self-Service Storage, Multi-Tenancy, and GitOps
Leaving vSphere retires more than a hypervisor. It retires the storage policy, DRS, and template model platform teams used to hand VMs to application teams. This post maps that VMware admin workflow onto Kubernetes-native equivalents: StorageClasses instead of storage policies, quotas instead of DRS limits, and GitOps-managed manifests instead of vCenter click-ops.
Rob Pankow
Aug 3rd, 2026 | 7 min read
Migrating VMware VMs to KubeVirt: A Storage-First Guide
Moving a VMware VM to KubeVirt starts with a storage decision, not a hypervisor one. Every disk becomes a PersistentVolumeClaim through a CDI DataVolume import, the live migration cutover window needs ReadWriteMany block storage, and a failed cutover should mean a snapshot restore, not a re-run of the conversion. This guide covers the storage mechanics most migration guides skip.
Rob Pankow
Aug 1st, 2026 | 8 min read
NATS JetStream on Kubernetes: Persistent Storage Architecture for Streaming Workloads
NATS JetStream embeds consensus and message persistence in the same server process, so a replicated stream's Raft log and its message store both land on whatever volume the node was given. Here is how the two I/O profiles differ, and how disaggregated NVMe/TCP and CSI snapshots fix JetStream storage on Kubernetes.
Rob Pankow
Jul 31st, 2026 | 7 min read
Apache Pulsar on Kubernetes: BookKeeper Storage Architecture and the Compute/Storage Split
Apache Pulsar already separates brokers from storage, but the BookKeeper bookie underneath still needs two distinct I/O profiles. Here is how journal and ledger storage differ, and how disaggregated NVMe/TCP and CSI snapshots fix Pulsar storage on Kubernetes.
Rob Pankow
Jul 30th, 2026 | 7 min read
OpenSearch on Kubernetes: Storage Architecture for Search and Observability Workloads
OpenSearch's shard-based segment merges compete with query reads on generic Kubernetes volumes. Here is why that I/O pattern breaks default StorageClasses, and how disaggregated NVMe/TCP and CSI snapshots fix OpenSearch storage on Kubernetes.
Rob Pankow
Jul 29th, 2026 | 7 min read
ClickHouse on Kubernetes: Storage Architecture for Analytics and Observability Workloads
ClickHouse's MergeTree engine writes immutable parts and merges them continuously in the background. Here is why that I/O pattern fights generic Kubernetes volumes, and how disaggregated NVMe/TCP and CSI snapshots fix ClickHouse storage on Kubernetes.
Rob Pankow
Jul 28th, 2026 | 7 min read
Zero-Copy VMware-to-HCI Migration: What It Doesn't Solve for Storage Architecture on Kubernetes
Hyperconverged infrastructure vendors are marketing zero-copy migration paths that convert VMware VVols directly into their own native virtual disk format, skipping a byte-for-byte data copy. That solves the one-time cost of the move. It does not solve the storage architecture question the migration creates: your data is now coupled to a new proprietary hypervisor instead of an old one.
Rob Pankow
Jul 27th, 2026 | 9 min read
Why CSI Alone Can't Absorb CRUD Churn at Scale: Provisioning Latency and Day 2 Operations
Every CSI provision, snapshot, and clone request is two calls, not one: a Kubernetes API object, then a separate round trip to an external array controller. At scale, that second hop becomes the bottleneck. Here is why CRUD churn breaks it, and what closes the Day 2 gap CSI leaves open.
Rob Pankow
Jul 23rd, 2026 | 8 min read
etcd Storage Architecture: Why Disk Latency Breaks Kubernetes Control Planes
Every object in a Kubernetes cluster, every pod, secret, and custom resource, lives in etcd, and every write to etcd must fsync to disk before the Raft log entry counts as committed. This guide covers etcd's Raft commit path, why disk tail latency causes leader elections and API slowdowns, and how disaggregated NVMe/TCP storage gives the control plane a consistent latency floor that local disks cannot always guarantee.
Rob Pankow
Jul 22nd, 2026 | 7 min read
Vector Database Storage on Kubernetes: Milvus Architecture for AI Workloads
Milvus splits vector data across growing segments, a write-ahead log, and object-storage-backed sealed segments, and its own docs point teams at S3 for all three. Here is why the hot tiers need disaggregated NVMe/TCP instead, and how CSI snapshots replace ad hoc collection backups.
Rob Pankow
Jul 21st, 2026 | 11 min read
CockroachDB on Kubernetes: Persistent Storage Architecture and High Availability
CockroachDB replicates data through per-range Raft consensus and stores it on disk with the Pebble LSM engine, so every write's commit latency and every node's recovery time are gated by the storage layer underneath. This guide covers persistent volume architecture for the CockroachDB Kubernetes Operator and Helm chart, range replication and rebalancing, and how disaggregated NVMe/TCP storage turns a full range up-replication into a volume reattachment.
Rob Pankow
Jul 20th, 2026 | 9 min read
AI Training Checkpoint and Model Registry Storage on Kubernetes
Training and fine-tuning jobs write large, bursty checkpoints that stall the GPU loop on the wrong storage backend, and most model registries are still ad hoc object-storage copies with no versioned rollback. Here is how disaggregated NVMe/TCP and CSI snapshots fix both.
Rob Pankow
Jul 17th, 2026 | 9 min read
SUSE Harvester Storage Architecture: Replacing Longhorn with Disaggregated NVMe/TCP
Harvester bundles Longhorn as its default storage engine, so every VM disk inherits Longhorn's per-volume replication overhead and full-replica rebuild time on node failure. This guide covers Harvester's storage architecture, what SUSE's new Certified Storage for Virtualization framework opens up, and how disaggregated NVMe/TCP changes the node-failure recovery model for VM workloads.
Rob Pankow
Jul 16th, 2026 | 11 min read
RabbitMQ on Kubernetes: Persistent Storage Architecture and High Availability
RabbitMQ's quorum queues replicate through a Raft log, and classic durable queues write to a per-node message store. Both paths depend on Kubernetes persistent storage in ways most CSI drivers were not designed around. This guide covers volume architecture for the RabbitMQ Cluster Operator, quorum queue replica recovery, and how disaggregated NVMe/TCP storage turns a full Raft snapshot resync into a volume reattachment.
Rob Pankow
Jul 15th, 2026 | 12 min read
Cassandra on Kubernetes: Persistent Storage Architecture and High Availability
Apache Cassandra's commit log and SSTable write path put specific demands on Kubernetes persistent storage. This guide covers volume architecture for cass-operator and K8ssandra, replication factor and consistency level design, and how disaggregated NVMe/TCP storage removes the full-stream rebuild penalty when a Cassandra node fails.
Rob Pankow
Jul 10th, 2026 | 10 min read
Kubernetes Multi-Tenant Storage Isolation: Per-Tenant NVMe/TCP After VMware Migration
Kubernetes namespace isolation controls CPU and memory but leaves storage unsolved. This guide covers how disaggregated NVMe/TCP delivers per-tenant QoS at the storage layer, the StorageClass design pattern, and the RBAC model that gives post-VMware teams the storage isolation they had in vSphere.
Rob Pankow
Jul 8th, 2026 | 11 min read
AI Inference Storage on Kubernetes: Why Model Weight Loading Needs NVMe/TCP
AI inference pods look stateless but behave like the most demanding stateful workloads in your cluster. Model weight loading, KV-cache persistence, and GPU-node locality pinning are storage problems. Here is why disaggregated NVMe/TCP solves them where hyperconverged storage cannot.
Rob Pankow
Jun 29th, 2026 | 11 min read
Day 2 Storage Operations After the VMware Exit: Snapshot, DR, and Observability on Disaggregated NVMe/TCP
The VMware migration plan covers what you are leaving. It rarely covers what you need to operate after the exit. vCenter handled snapshots, replication, and storage observability automatically. Kubernetes does not. This guide walks through how disaggregated NVMe/TCP delivers all three as native capabilities of the storage layer, without an overlay.
Rob Pankow
Jun 26th, 2026 | 11 min read
After the VMware Exit: Replacing Overlay Storage on Kubernetes with Disaggregated NVMe/TCP
Most VMware migration guides cover what you're leaving. Almost none address the storage decision on the other side. Teams that migrate to Kubernetes without rethinking storage end up rebuilding the same hyper-converged coupling problem they escaped — just in container form. This guide explains the overlay storage tax, why it replicates vSAN's failure modes, and how disaggregated NVMe/TCP breaks the pattern.
Rob Pankow
Jun 25th, 2026 | 10 min read
vSAN Exit Cost Economics: Broadcom Licensing vs. Disaggregated NVMe/TCP
Broadcom's post-acquisition pricing reset pushed vSAN licensing from a predictable line item to a per-core contract that scales with every server refresh. This guide models the real cost of VMware vSAN at 200 cores under VCF and VVF, compares it to commodity NVMe/TCP storage, and explains what the per-core number actually omits.
Rob Pankow
Jun 24th, 2026 | 13 min read
Kafka on Kubernetes: Persistent Storage Architecture and High Availability
Running Apache Kafka on Kubernetes couples broker pods to the nodes where their log data lives. This guide covers persistent volume architecture, StorageClass configuration for Strimzi and Confluent Operator, and how disaggregated NVMe/TCP storage removes broker pod pinning to enable zero-downtime rolling upgrades and true HA.
Rob Pankow
Jun 23rd, 2026 | 12 min read
OpenShift Virtualization Storage Architecture: Disaggregated NVMe for Live Migration and Multi-Tenancy
OpenShift Virtualization requires ReadWriteMany block storage for VM live migration and CSI-accessible persistent volumes for tenant Kubernetes clusters running as VMs. This guide covers the storage architecture options and how disaggregated NVMe/TCP meets both requirements without a proprietary array dependency.
Rob Pankow
Jun 22nd, 2026 | 11 min read
Redis on Kubernetes: Persistent Storage Architecture and High Availability
Running Redis on Kubernetes requires storage that handles AOF fsync writes, RDB snapshot I/O bursts, and per-replica PVC constraints. This guide covers persistent volume architecture, StorageClass configuration, and how disaggregated NVMe/TCP storage removes node-affinity pinning for Redis Sentinel and Redis Cluster deployments.
Rob Pankow
Jun 19th, 2026 | 9 min read
MongoDB on Kubernetes: Persistent Storage Architecture and High Availability
Running MongoDB on Kubernetes requires storage that handles WiredTiger's concurrent journal writes, oplog replication, and per-replica PVC constraints. This guide covers persistent volume architecture, StorageClass configuration, and how disaggregated NVMe/TCP storage removes node-affinity pinning for high availability MongoDB replica sets.
Rob Pankow
Jun 18th, 2026 | 9 min read
MySQL on Kubernetes: Persistent Storage Architecture and High Availability
Running MySQL on Kubernetes requires storage that handles InnoDB's small-block random writes, redo log fsync calls, and replica scheduling constraints. This guide covers persistent volume architecture, StorageClass configuration, and how disaggregated NVMe/TCP storage removes node-affinity pinning for high availability MySQL clusters.
Rob Pankow
Jun 16th, 2026 | 11 min read
Encrypting Kubernetes Persistent Volumes with Vault KMS
Application-level encryption and host-disk encryption both leave gaps in Kubernetes environments. Storage-layer encryption with Vault KMS encrypts NVMe/TCP persistent volumes at the storage system, meeting PCI DSS and SOC 2 requirements without any application changes.
Rob Pankow
Jun 16th, 2026 | 7 min read
Storage for Neoclouds: The Data Layer GPU Clouds Cannot Ignore
Neoclouds win or lose on GPU utilization, and idle GPUs are almost always a storage problem. Here is how simplyblock fits as the sovereign, cost-efficient, software-defined storage layer for GPU clouds.
Rob Pankow
Jun 16th, 2026 | 11 min read
Top 5 European Storage Solutions in 2026
The best European storage solutions for sovereign, self-hosted infrastructure in 2026, with an open-source focus: simplyblock, Longhorn, LINBIT, Ceph, and MooseFS, plus a clear way to choose.
Rob Pankow
Jun 15th, 2026 | 8 min read
Elasticsearch on Kubernetes: Persistent Storage Architecture for Production
Elasticsearch on Kubernetes demands careful storage design: hot-tier nodes need 10,000+ IOPS for indexing and merge, while warm and cold tiers have very different requirements. This guide covers how to size and configure persistent volumes for each data tier, and how disaggregated NVMe/TCP storage removes the local-disk scheduling constraint.
Rob Pankow
Jun 15th, 2026 | 9 min read
KubeVirt Persistent Storage: Architecture Guide for Production VM Workloads
KubeVirt VMs on Kubernetes are backed by PVCs, which means live migration and fault tolerance depend entirely on your storage model. This guide covers how disaggregated NVMe/TCP block storage removes the local-disk scheduling constraint and enables unrestricted VM live migration on Kubernetes.
Rob Pankow
Jun 14th, 2026 | 7 min read
The European Data Infrastructure Stack in 2026
A practical, layer-by-layer guide to building a sovereign European data infrastructure stack in 2026: compute, orchestration, storage, and data, with simplyblock as the storage foundation.
Rob Pankow
Jun 12th, 2026 | 9 min read
NVMe/TCP vs NVMe/RoCE for Kubernetes Storage: Choosing the Right Fabric
NVMe over Fabrics gives Kubernetes clusters low-latency block storage over the network. The transport you pick, TCP or RoCE, determines your latency floor, infrastructure cost, and operational complexity. Here is how to choose.
Rob Pankow
Jun 12th, 2026 | 9 min read
Why CSI Alone Isn't Enough for Stateful Kubernetes
The Container Storage Interface connects storage to pods one volume at a time. That 1:1 model is fine for small clusters and quietly breaks at scale. Here is why, and how a pooled NVMe fabric fixes it.
Rob Pankow
Jun 8th, 2026 | 11 min read
How to Choose Persistent Storage for Kubernetes: A Decision Framework
Choosing Kubernetes persistent storage is a multi-dimensional decision: performance, data protection, cost, operations, and compliance all point in different directions. This framework gives platform engineers a structured way to evaluate storage options against the workloads they actually run.
Rob Pankow
Jun 3rd, 2026 | 10 min read
NVMe Storage Cost Optimization in 2026: Erasure Coding, Thin Provisioning, and Compute Efficiency
NVMe drives deliver the performance Kubernetes stateful workloads need, but triple replication and thick provisioning multiply their cost fast. Here is a practical breakdown of erasure coding economics, thin provisioning, and how sub-millisecond latency reduces compute waste.
Rob Pankow
May 26th, 2026 | 9 min read
Storage Testing for Fintech and Healthcare: Why Generic QA Is Not Enough
Platform engineers at regulated companies face a version of the storage reliability question that goes beyond typical availability metrics. Generic QA tests that the system recovers. Regulated workloads require a more specific answer: that the system behaves correctly under the exact failure modes that correspond to specific regulatory obligations.
Rob Pankow
Apr 6th, 2026 | 4 min read
AI Workloads on OpenShift: Best Practices
Learn best practices for running AI workloads on OpenShift, from GPU utilization and data locality to storage performance, checkpoints, and private-cloud operations.
Rob Pankow
Apr 6th, 2026 | 4 min read
Databases on OpenShift: Best Practices
Learn best practices for running databases on OpenShift, including storage classes, topology, backup, branching, QoS, and platform ownership.
Rob Pankow
Apr 6th, 2026 | 4 min read
Unified Storage for Containers and Virtual Machines on OpenShift
Learn how to design unified OpenShift storage for containers and virtual machines without forcing VM workloads and Kubernetes-native applications into separate storage models.
Rob Pankow
Apr 3rd, 2026 | 6 min read
The 2026 OpenShift Data Stack
The 2026 OpenShift data stack is not just containers plus storage. It combines virtualization, Kubernetes-native block storage, and a better database workflow for private-cloud and VMware-exit programs.
Rob Pankow
Mar 29th, 2026 | 4 min read
How to Do Storage Disaster Recovery with OpenShift
A technical guide to designing storage disaster recovery on OpenShift, including RPO/RTO targets, replication patterns, failover runbooks, and recovery validation.
Rob Pankow
Mar 28th, 2026 | 8 min read
HCI Storage for OpenShift: Software-Defined Flexibility vs. Appliance Lock-In
Appliance-based HCI bundles compute, storage, and OpenShift into a single vendor stack — but the hardware lock-in and premium cost often outweigh the convenience. Software-defined HCI gives platform teams the same operational model on commodity NVMe hardware.
Rob Pankow
Mar 28th, 2026 | 8 min read
OpenShift Data Foundation Alternative: What to Evaluate Before Committing to ODF
ODF is the default Red Hat storage recommendation for OpenShift, but it inherits Ceph's operational complexity and performance tradeoffs. Here is what platform teams should evaluate before committing.
Rob Pankow
Mar 28th, 2026 | 8 min read
Replacing vSAN Storage When Moving to OpenShift or Kubernetes
Most VMware migration guides focus on the compute layer — but vSAN is the storage layer, and it does not migrate. Here is what platform teams need to plan when replacing vSAN as part of an OpenShift or Kubernetes transition.
Rob Pankow
Mar 27th, 2026 | 8 min read
Best Openshift HCI Storage Solutions 2026
The best OpenShift HCI storage solutions in 2026 depend on workload criticality, platform maturity, and operating model. Compare simplyblock, OpenShift Data Foundation, and Nutanix.
Rob Pankow
Mar 24th, 2026 | 9 min read
Volume Mobility Across Zones and Clusters for Stateful Kubernetes
Stateful Kubernetes reliability depends on volume mobility strategy across zones and clusters, including failover design, restore pathways, and operational controls.
Rob Pankow
Mar 20th, 2026 | 8 min read
Private Cloud Storage for Proxmox and Kubernetes
A unified storage foundation for Proxmox and Kubernetes helps enterprise teams modernize private cloud estates without splitting operations across VM and container silos.
Rob Pankow
Mar 18th, 2026 | 9 min read
Ransomware Recovery for Kubernetes Storage
Ransomware recovery in Kubernetes depends on immutable restore points, clean recovery workflows, and storage architecture that limits blast radius under pressure.
Rob Pankow
Mar 16th, 2026 | 9 min read
Backup and Restore for KubeVirt and VM Workloads on Kubernetes
Platform teams running containers and VMs together need unified backup and restore workflows that protect both workload types on one storage foundation. This post walks through the tooling, consistency challenges, and RPO/RTO design that make it work.
Rob Pankow
Mar 12th, 2026 | 5 min read
How to make Storage Agentic Ready?
Making storage agentic ready in 2026 means designing for low-latency context retrieval, concurrent state updates, and operational resilience under autonomous multi-step workloads.
Rob Pankow
Mar 12th, 2026 | 8 min read
How to Design a Kubernetes StorageClass for PostgreSQL Performance
Most PostgreSQL issues blamed on Kubernetes are really storage-class mistakes. Here is how to set replication, topology, and expansion for predictable performance.
Rob Pankow
Mar 10th, 2026 | 10 min read
Application-Consistent Backups for PostgreSQL on Kubernetes
Reliable PostgreSQL recovery on Kubernetes requires more than storage snapshots. Teams need clear consistency strategy across snapshots, WAL, and restore workflows. This post covers the full model — from checkpoint semantics to PITR and operator integration.
Rob Pankow
Mar 7th, 2026 | 6 min read
Best Database Storage 2026
The best database storage in 2026 depends on latency consistency, operational model, and scaling trajectory. Compare Simplyblock, Amazon EBS io2 Block Express, and Ceph.
Rob Pankow
Mar 6th, 2026 | 6 min read
Best Postgres Storage in 2026
The best Postgres storage in 2026 depends on latency targets, operational model, and scaling plans. Compare Simplyblock, Amazon EBS io2 Block Express, and Ceph.
Rob Pankow
Mar 5th, 2026 | 4 min read
Simplyblock vs vSAN
Compare Simplyblock vs vSAN for modern stateful infrastructure: architecture, performance, operations, and migration fit for Kubernetes and OpenShift programs.
Rob Pankow
Mar 1st, 2026 | 6 min read
Best Rancher Storage 2026
The best Rancher storage choice in 2026 depends on workload profile and operational capacity. Compare simplyblock, OpenEBS, and Ceph with practical tradeoffs.
Rob Pankow
Feb 27th, 2026 | 8 min read
Best Kubernetes Storage 2026
The best Kubernetes storage in 2026 depends on workload profile, operations model, and performance targets. Compare simplyblock, OpenEBS, and Ceph with practical tradeoffs.
Rob Pankow
Feb 19th, 2026 | 5 min read
Simplyblock vs. OpenShift Data Foundation: A Technical Storage Comparison
A technical comparison of simplyblock and OpenShift Data Foundation for OpenShift teams choosing between NVMe-native block storage and Ceph-based integrated storage.
Rob Pankow
Jan 12th, 2026 | 11 min read
DPUs for Storage: What They Are, and Why They Matter
Data centers changed faster than most storage stacks. NVMe SSDs became standard. Ethernet became faster and cheaper. Kubernetes became the default platform for new infrastructure. At the same time,…
Rob Pankow
Dec 12th, 2025 | 8 min read
Kubernetes Storage: Disaggregated or Hyper-converged?
Modern cloud-native environments demand more from storage than ever before. As Kubernetes becomes the dominant platform for deploying applications at scale, teams are confronted with a critical…
Rob Pankow
Dec 10th, 2025 | 10 min read
NVMe/TCP vs NVMe/RoCE: Which Protocol For High-Performance Storage?
As modern workloads become faster, smarter, and more distributed, the infrastructure behind them must keep up. Enterprise applications, especially those driven by AI, analytics, and cloud-native…
Rob Pankow
Apr 15th, 2025 | 6 min read
Cloud Exit Is Real: Why Cloud Economics Break Down at Scale
When I wrote “Why Companies Are Ditching the Cloud: The Rise of Cloud Repatriation” last year, the idea of moving workloads off the public cloud still felt somewhat radical in mainstream…
Rob Pankow
Feb 26th, 2025 | 5 min read
Top 5 Open Source Software-Defined Storage (SDS) Tools for Enterprises
📅 Updated for 2025 Read Time: 6 min Enterprises today generate and manage petabytes of data, with storage demands growing exponentially. Traditional hardware-based storage solutions lack…
Rob Pankow
Feb 20th, 2025 | 11 min read
Alternative to SAN - Why Software-Defined Storage Wins in the Long Run
NetApp has been a dominant player in the enterprise storage market for over two decades, offering feature-rich SAN and NAS solutions but it's time for an alternative. The landscape of enterprise…
Rob Pankow
Feb 20th, 2025 | 9 min read
Alternative to SAN - Why Software-Defined Storage Wins in the Long Run
SAN systems have been dominant in the enterprise storage market for over two decades. The landscape of enterprise storage is changing. Enterprises are no longer just evaluating storage solutions…
Rob Pankow
Feb 4th, 2025 | 9 min read
How To Reduce Your Cloud Storage Carbon Footprint
Discover practical strategies for data center carbon footprint reduction and storage optimization in the cloud infrastructure.
Rob Pankow
Jan 28th, 2025 | 7 min read
Bare-Metal Kubernetes - Power of Direct Hardware Access
As of 2025, over 90% of enterprises have adopted Kubernetes. The technology has matured, and the ecosystem around containerization technologies is richer than ever. More platform teams are realizing…
Rob Pankow
Jan 21st, 2025 | 9 min read
Database Performance: Impact of Storage Limitations
TLDR: Storage and storage limitations have a fundamental impact on database performance, with access latency creating a hard physical limitation on IOPS, queries per second (QPS), and transactions…
Rob Pankow
Dec 23rd, 2024 | 10 min read
Kubernetes Storage 201: Concepts and Practical Examples
Kubernetes storage is a sophisticated ecosystem designed to address the complex data management needs of containerized applications. At its core, Kubernetes storage provides a flexible mechanism to…
Rob Pankow
Oct 17th, 2024 | 12 min read
NVMe Storage for Database Optimization: Lessons from Tech Giants
Leveraging NVMe-based storage for databases brings whole new set of capabilities and performance optimization opportunities. In this blog we explore how can you adopt NVMe storage for your database workloads with case studies from tech giants such as Pinterest or Discord.
Rob Pankow
Oct 3rd, 2024 | 4 min read
Local NVMe Storage on AWS - Pros and Cons
The debate over the optimal storage solution has been ongoing. Local instance storage on AWS (i.e. ephemeral NVMe disk attached to EC2 instance) brings remarkable cost-performance ratios. It offers…
Rob Pankow
Sep 12th, 2024 | 7 min read
AWS Migration - How to Migrate into the Cloud? Data Storage Perspective
Migrating to the cloud can be daunting, but it becomes a manageable and rewarding process with the right approach and understanding of the storage perspective. Amazon Web Services (AWS) offers a…
Rob Pankow
Sep 12th, 2024 | 14 min read
RDS vs. EKS: The True Cost of Database Management
Databases can make up a significant portion of the costs for a variety of businesses and enterprises, and in particular for SaaS, Fintech, or E-commerce & Retail verticals. Choosing the right…
Rob Pankow
Sep 12th, 2024 | 6 min read
What is the AWS Workload Migration Program and how simplyblock can help you with cloud migration?
The AWS Workload Migration Program is a comprehensive framework designed to help organizations migrate their workloads to the AWS cloud efficiently and effectively. It encompasses a range of tools,…
Rob Pankow
Sep 6th, 2024 | 7 min read
Amazon EKS vs. ECS: Understanding the Differences and Choosing the Right Service
When it comes to container orchestration on AWS, two primary services come to mind: Amazon Elastic Kubernetes Service (EKS) and Amazon Elastic Container Service (ECS) . Both offer robust solutions…
Rob Pankow
Aug 26th, 2024 | 7 min read
Avoiding Storage Lock-in - Block Storage Migration with Simplyblock
Storage and particularly block storage is generally easy to migrate. It doesn’t create vendor lock-in, which is very different from most database systems. Therefore, it’s worth to briefly line out…
Rob Pankow
Aug 11th, 2024 | 11 min read
AWS Storage Optimization: Best Practices for Cost and Performance
Managing storage costs in AWS environments has become increasingly critical as organizations scale their cloud infrastructure. With storage often representing 20-30% of cloud spending, AWS storage…
Rob Pankow
Aug 6th, 2024 | 9 min read
What is AWS Marketplace?
Overview of AWS Marketplace AWS Marketplace is a digital catalog that allows customers to find, buy, and deploy software and services that run on Amazon Web Services (AWS) . It's designed to simplify…
Rob Pankow
Jul 23rd, 2024 | 6 min read
Rockset alternatives - migrate with simplyblock
On June 21, 2024, Rockset announced its acquisition by OpenAI , setting off a countdown for many organizations using their database. If you're a Rockset user, you're now facing a critical deadline:…
Rob Pankow
Jul 22nd, 2024 | 8 min read
What is the AWS Savings Plan? Guide to AWS Discount Programs
As cloud computing continues to transform how businesses operate, Amazon Web Services (AWS) remains the largest cloud provider globally, offering scalable and cost-effective solutions for…
Rob Pankow
Jun 28th, 2024 | 10 min read
How to reduce AWS cloud costs with AWS marketplace products?
The AWS Marketplace is a comprehensive catalog consisting of thousands of offerings that help organizations find, purchase, deploy and manage third-party software and services to optimize their cloud…
Rob Pankow
Jun 18th, 2024 | 11 min read
What are AWS Credits and how to get them?
AWS Credits are cash credits provided by Amazon Web Services (AWS) to help offset the costs of using AWS cloud services. These credits can be applied to a wide range of AWS services, such as compute,…
Rob Pankow
Jun 13th, 2024 | 7 min read
How to benefit from AWS Enterprise Discount Program (EDP)
The AWS Enterprise Discount Program (EDP) is a discount initiative designed for organizations spending at least $1M per year on AWS cloud services and committed to extensive and long-term usage of…
Rob Pankow
Apr 30th, 2024 | 11 min read
AI Storage: How To Build Scalable Data Infrastructures for AI workloads?
AI workloads bring new requirements to your AI storage infrastructure, marking a significant change compared to the “ML era” of Big Data storage. The average scale of an AI dataset is multiple times…
Rob Pankow
Apr 17th, 2024 | 7 min read
Block Storage Volume Pooling for the Cloud-Age
If you have services running in the AWS, you’ll eventually need block storage to store data. Services like Amazon EBS (Elastic Block Storage) provide block storage to be used in your EC2 instances,…
Rob Pankow
Mar 13th, 2024 | 7 min read
The Next-Gen of Storage Virtualization for Modern Cloud Computing
Storage virtualization, as a concept, has been around for about 60 years now. However, it became more prominent and practical with the advancement of computing technologies in the late 1990s and…
Rob Pankow
Apr 5th, 2023 | 3 min read
Simplyblock for AWS - Performance Test
Simplyblock's NVMe-oF storage on AWS delivers >128 Gbit/s throughput, 3 GiB/s per GPU, and sub-100µs latency — validated in MLPerf Storage v2.0 and ClickBench benchmarks.
Rob Pankow
Apr 5th, 2023 | 2 min read
Simplyblock for AWS (demo videos)
We have recently conducted a demo showcasing capabilities of simplyblock's high-perfromance block storage on the AWS platform. The demo offers a glimpse into the seamless integration, exceptional…
Rob Pankow
Apr 5th, 2023 | 2 min read
Simplyblock for AWS - Whitepaper
Simplyblock has recognized the limitations of conventional cloud storage services, specifically AWS Elastic Block Storage (EBS) , and has introduced a groundbreaking alternative. This whitepaper…
Rob Pankow
Feb 27th, 2023 | 4 min read
Reducing cloud costs by 30%: a case study on relational databases with hybrid cloud
The benefits of a hybrid cloud operating model are obvious: take the best of both worlds . Flexibility, fast time-to-market, cost-effective high availability and a huge variety of great platform…