TM

Solutions

Consolidate without levelling down.

True unified block and file in one unit, with each workload carrying its own profile, block size, cache behaviour and protection topology.

Book a demoSee the architecture
  • Unified block and file
  • Per-workload profiles
  • Immutable snapshots

The problem

Consolidation usually means everyone gets the average.

The business case for consolidation is straightforward: fewer systems, fewer support contracts, fewer things to learn. The catch appears at configuration time, when block size, cache policy and protection level are set for the array rather than for the workload.

So the database gets settings tuned for a file share, the file share carries overhead it does not need, and the virtualisation estate sits somewhere in between. Everyone gets the average, and the workload that mattered most is the one that notices.

The usual escape is to buy separate systems again — at which point the consolidation savings have been spent on avoiding the consequence of consolidating.

Conventional array

one setting, applied to everything

Array-wide block size — a single compromise value
  • VM datastoreSmall block, low latency
  • Database volumesSmall block, no dedup
  • File sharesLarge block, deduplicated

The bar is what every workload is given; the tick is what it needed. The gap between them is the compromise, and the workload that mattered most is the one that notices.

AiRE — one unit

a setting per data set

  • VM datastoreSmall block, low latency
  • Database volumesSmall block, no dedup
  • File sharesLarge block, deduplicated

Each data set carries its own block size, cache behaviour, reduction and protection topology — on the same system.

Bar length represents block size. The consolidation trade-off is not about capacity — it is about whether unlike workloads are forced onto one setting.

What STORViX does about it

Different settings for different workloads, same unit.

AiRE manages both block and file protocols within the same unit, and configuration lives at the data set. That is the whole answer to the consolidation trade-off.

  • The Performance profile for latency-sensitive work

    Virtual machines and transactional databases get a profile that prioritises lower latency and higher throughput, with policies tuning I/O path management and cache behaviour to match.

    • Lower latency and higher throughput
    • Suited to VM datastores
    • Policies adjustable from best-practice defaults

  • Variable block sizes per data set

    A database and a file share want different block sizes. AiRE supports variable block sizes for volumes and file systems, so neither has to accept the other's setting.

    • Per-volume and per-filesystem
    • Set from the workload profile
    • No array-wide compromise

  • The Resilience profile for what cannot be lost

    Mission-critical data sets get redundant backups, external replication and protection mechanisms weighted toward preventing loss rather than toward speed.

    • Redundant backup
    • External replication
    • RAID-Z or mirror topologies

  • Recovery measured in a revert, not a restore

    Immutable copy-on-write snapshots allow a data set to be returned to a known-good point in time. For a corrupted database or an encrypted volume, that is a materially different recovery path from a tape restore.

    • Point-in-time immutable snapshots
    • Instant clones for test restores
    • Block-level incremental off-site copies

  • Non-disruptive capacity growth

    Capacity is added in 4U disk expansion units without downtime. A critical application does not need an outage window because the estate grew.

    • Upgradable with zero downtime
    • Modular expansion
    • No forklift replacement

  • Automated response before a human notices

    CloudSight raises AI-driven alerts on emerging problems, and AutoPILOT resolves what it can autonomously — escalating what it cannot rather than failing quietly.

    • Real-time telemetry analysis
    • Self-healing algorithms
    • Escalation to SmartCARE

Outcomes

What changes in practice

  • Consolidation without a performance argument

    Each workload keeps its own settings, so the case for consolidating is not undermined by the first application that regresses.

  • A published sub-millisecond result

    Eteria achieved sub-millisecond maximum latency on an all-flash AiRE configuration integrated with Zerto, while keeping hot copies of customer VMs available continuously.

  • Recovery that has actually been rehearsed

    Instant, space-free clones make testing a restore cheap enough to do regularly rather than annually.

  • Growth without outage windows

    Expansion is non-disruptive, so capacity planning stops being constrained by change-freeze calendars.

Run the demo against your workload, not ours.

Bring the workload mix you are consolidating and the latency requirement that matters. A demo that ignores those is not worth the hour.

Book a demoTalk to an architect