HCIandHDD

by

On-Demand · No Registration

How to Use HCI and Hard Drives to Fight Data Center Inflation

Watch the whole session, with every demo. No registration required.

Watch Now

Hard drive tiering moved back to the top of my list the day Dave Vincent and I priced drives to expand our lab. A 15.36 TB NVMe SSD quoted at $10,850, or about $706 per terabyte. A 24 TB Western Digital hard drive came in at $1,350, about $56 per terabyte. Flash now costs more than 12 times as much per terabyte as disk, and on October 7 Dave and I spent an hour with a record audience working through what IT teams can do about it.

Watch the whole session, with every demo, in the on-demand presentation. No registration required.

Key Takeaways
  • Data center inflation now reaches the drive bay, and putting every workload on flash has become an expensive default.
  • VergeOS sets the tier per virtual drive and moves it live, so the administrator decides placement and can reverse it in seconds.
  • Metadata on NVMe, write reordering, read prefetch and a deduplicated RAM cache make hard drives viable for the long tail of workloads.

Data Center Inflation Has Reached the Drive Bay

I love flash as much as the next guy, but I don’t love it 12 times as much. The price gap keeps widening, and I see no relief coming soon. A new fab takes at least two years to build, longer in countries with strict permitting, and the AI factories driving this demand will absorb new supply as fast as it ships.

Raw media cost, from one terabyte to a 150 TB estate, NVMe flash versus hard drives One terabyte costs about 706 dollars on NVMe flash and about 56 dollars on hard drives. Across a 150 TB estate that grows to 105,900 dollars on flash against 8,400 dollars on hard drives, 97,500 dollars less in raw media. One terabyte, then a 150 TB estate 1 TB of capacity 150 TB of capacity Raw media cost per terabyte Raw media cost for 150 TB All flash · 15.36 TB NVMe $706 $105,900 Hard drives · 24 TB HDD $56 $8,400 $97,500 less in raw media
Kioxia 15.36 TB NVMe at $10,850 against a WD 24 TB hard drive at $1,350, checked September 28, 2026. Raw media for a single copy, before redundancy.

When flash got cheaper every quarter, putting everything on it was like hitting a watermelon with a sledgehammer. It worked, and nobody questioned the bill. Back in my analyst days, friends at the all-flash array vendors told me their systems averaged under 5,000 IOPS. Most data sits idle most of the time. At today’s prices it pays to slice the watermelon, which means IT takes control of placement again.

Pick Your Tier, Then Change Your Mind

The first generation of hybrid storage relied on auto-tiering, where an algorithm decided which blocks went to disk. It moved data you wanted on flash, and undoing its choices meant scripting a touch of every file. I covered why that failed in Hybrid Storage Failed Once. Will It Fail Again? on StorageSwiss.

VergeOS hands the decision to the administrator. Each virtual drive has a preferred tier, and changing it starts a live, background migration with no reboot and no maintenance window. During the webinar Dave ran a SQL Server VM processing about 80,000 rows per second, moved its data drive from Tier 1 NVMe to Tier 5 hard drives, and moved it back. The query monitor never flinched. From the CLI, the same change took about 1.3 seconds to accept.

Hard drive tiering in VergeOS: moving a live SQL Server drive between NVMe and hard drives A SQL Server VM keeps its OS, transaction log and tempdb drives on Tier 1 NVMe. Its data drive moves from Tier 1 NVMe to Tier 5 hard drives and back while queries keep running. Preferred tier, set per virtual drive and changed live SQL Server VM OS · Tier 1 NVMe Transaction log · Tier 1 tempdb · Tier 1 NVMe Data · Tier 1 NVMe Data · Tier 5 hard drives Tier 1 · NVMe flash Tier 5 · hard drives SQL data drive Query monitor: about 80,000 rows per second, steady through both moves
The data drive moves to hard drives and back while the OS, log and tempdb stay on NVMe. The VM never stops.

That speed is what makes hard drive tiering practical. You can be aggressive, push the long tail to disk, and move a drive back to flash the moment someone complains. Dave also showed a local AI agent finding idle drives and moving them with the same CLI. If you want the cost math on a single database VM, Dave walks through it in Can HCI Run a Database Application on Hard Disk Drives?

How VergeOS Makes Hard Drives Perform

Our CTO, Greg Campbell, built most of this a decade ago, right before the industry went all-flash. VergeOS keeps all metadata on Tier 0 NVMe, so lookups never touch a spinning disk. It reorders incoming writes from many VMs into sequential streams, which is how hard drives like to work, and prefetches reads across the cluster.

How VergeOS makes hard drives perform Random writes from several VMs enter VergeOS, which sends metadata to Tier 0 NVMe and reorders the data into a sequential stream for the hard drive tier. A deduplicated RAM cache on each host serves repeat reads from memory. Random writes in, sequential writes out VM 1 VM 2 VM 3 VergeOS write reordering and read prefetch Tier 0 NVMe all metadata Hard drive tier sequential writes Host RAM cache deduplicated, next to the VM repeat reads served from memory
Metadata goes to Tier 0 NVMe, and writes from many VMs reach the hard drives as ordered, sequential streams.

Each host also keeps a RAM cache next to the VM. Global inline deduplication runs in the core code, so the cache holds no duplicate blocks and serves more of the working set from memory. Some of the performance in Dave’s demo came from that cache, and that is by design. It is why the disk tier sees far less random I/O than its raw specs suggest.

Protecting High-Capacity Drives

Hard drives fail, and so do SSDs, but a large hard drive takes longer to rebuild. VergeOS sets redundancy per system and applies it per tier. N+1 keeps two copies of every block and survives one failure. N+2 keeps three copies and survives two. A failure in the hard drive tier stays in that tier.

Frequent, deduplicated snapshots cover logical mistakes. For failures beyond the redundancy level, ioGuardian turns your DR site into a repair server. VergeOS pulls back the 64K blocks it needs from the remote copy, which avoids a rollback to a snapshot. The cost is latency on those reads. You can place a repair server in the primary data center to cut that latency, at the price of another set of drives.

ioGuardian supplies missing blocks from the DR site when drive failures exceed redundancy Three VMs each read from their own hard drive at the primary site. Two of the three drives fail, more than N+1 redundancy covers. The ioGuardian repair server at the DR site then supplies the missing 64K blocks to the two affected VMs, which keep running with added latency on those reads. ioGuardian: the DR copy doubles as a repair server Running VMs Primary site · hard drive tier, N+1 DR site · ioGuardian VM 1 Drive 1 Copy 1 VM 2 Drive 2 Copy 2 VM 3 Drive 3 Copy 3 Normal operation: each VM reads from its own drive 64K blocks Drives 2 and 3 fail, more than N+1 redundancy covers ioGuardian supplies the missing blocks from the DR site
Each VM reads from its own drive. When two drives fail, more than N+1 covers, ioGuardian at the DR site supplies the missing blocks, which avoids a rollback to a snapshot. The trade-off is latency on those reads.

What Attendees Asked

What happens on a cache miss? The read goes to the hard drive at its native speed. The sequential layout gives the drive the best possible access pattern, and the RAM cache absorbs the burst of requests, so the disk avoids hundreds of workloads hammering it at random.

How does VergeOS compare to Proxmox with Ceph? VergeOS is a single code base. You load one image and get compute, storage and networking in one product, which makes it faster to learn and run. Topgolf runs VergeOS across more than 100 venues, offices and data centers.

What server do I need for hard drives? Pick a chassis that fits the workload. That usually means a 2U server with more drive bays instead of a 1U flash-only node.

Where to Start

Pull a 30-day I/O report, find the VMs that sit idle, and set their preferred tier to hard drives. If a workload slows down, move it back.

The first step in fighting data center inflation is to avoid buying new hardware at all. VergeOS runs on the servers you already own, which also makes it the fastest path off VMware. Watch the on-demand presentation for every demo, no registration required, or book a technical whiteboard session and we will map your workloads to tiers.

Key Terms

Hard drive tiering

Placing data on hard disk drives instead of flash according to how often it is accessed.

Preferred tier

The VergeOS setting on each virtual drive that names the storage tier it should live on.

Long tail

The many workloads that generate little I/O over a 30-day period.

ioGuardian

A VergeOS repair server, usually at the DR site, that supplies missing blocks when failures exceed the redundancy level.

Frequently Asked Questions

Does moving a drive to a different tier require downtime?

The move runs in the background and the VM keeps running. Dave moved a live SQL Server data drive from NVMe to hard drives and back during the webinar.

Does VergeOS use auto-tiering?

The administrator sets the preferred tier for each virtual drive and can change it at any time, which keeps placement predictable and reversible.

Which workloads belong on hard drives?

Workloads in the long tail of a 30-day I/O report, such as file servers, ISO libraries, dev and test, backup targets and database data files with modest I/O.

Where can I watch the webinar?

The on-demand presentation includes every demo and requires no registration.

Filed Under: Storage Tagged With: , , , , , ,

by

A database VM is five workloads sharing one box. Put each drive on the tier it needs, and most of the capacity lands on hard disk drives.

Here’s the result, up front: I moved a live SQL Server data drive from NVMe onto hard disk drives. At today’s street prices, that one change cuts the VM’s storage media cost from $263 to $133. No downtime, no maintenance window. One setting on one drive, while the database kept answering queries.

Now imagine that across a whole data center estate. Take 200 TB: if even 75% of it can move from NVMe to hard disk drives, that’s 150 TB at $650 less per terabyte, or $97,500 in raw media saved. Even if you keep a second copy on ioGuardian for better than RF3 redundancy, that’s still more than $50,000 in savings.

On-Demand · No Registration

How to Use HCI and Hard Drives to Fight Data Center Inflation

George Crump and Dave Vincent move a running VM from flash to hard drives and back. Every demo included.

Watch Now

Why it matters now: for years the default answer to “which storage?” was “more flash,” because flash got cheaper every quarter. That era is over. At late-September street prices, a 15.36 TB NVMe SSD costs $706 per terabyte and a 24 TB hard drive costs $56, and flash prices keep climbing. I don’t see that reversing. Put everything on flash, just in case has become an expensive habit.

So the real question isn’t can a database run on hard disk drives. It’s what level of performance does each part of that database actually need, and what are you willing to pay for it? Nobody can answer that for your workload but you, and you’ll get there by experimenting. What VergeOS gives you is the ability to put each workload on the right tier, and move it again whenever the numbers, or the business, tell you to.

Key Takeaways
  • Flash prices are rising with no relief in sight, so “all flash, just in case” is now an expensive default.
  • A database VM is several workloads, and each one should get the level of performance it needs, not more.
  • The right tier is something you find by experimenting. On VergeOS, Preferred Tier is set per VM drive and changes move live, so every experiment is reversible and placement can change whenever business needs do.
  • In one example, moving a 200 GB SQL Server data drive to hard disk drives cut the VM’s media cost roughly in half.

A Database Is Several Workloads Wearing One VM

Take the example VM for the math: SQL Server, five drives, 510 GB total. Look at what each drive actually does and the “databases need flash” rule starts to fall apart.

DriveSizeWhat it does all dayWhere it belongs
OS80 GBBoot, patches, logsNVMe
Transaction log40 GBSmall sequential writes; every commit waits on itNVMe
tempdb40 GBBursty random I/O from sorts, joins, temp tablesNVMe
Data files200 GBCheckpoint writes, reads of cold pagesIt depends (often HDD)
Local backups150 GBLarge sequential streamsHDD
A SQL Server VM is five workloads with different I/O patterns Animated diagram of five SQL Server VM drives. The OS drive sees light I/O, the transaction log a steady stream of small sequential writes, and tempdb random bursts, so all three belong on NVMe. Most reads headed for the data drive stop at the buffer pool in RAM, so the data drive is a hard disk drive candidate. Backups are large sequential streams that suit hard disk drives. DriveI/O patternWhere it belongs OS80 GB NVMe Transaction log40 GB NVMe tempdb40 GB NVMe Data files200 GB HDD candidate Local backups150 GB HDD Buffer pool(RAM)
Three drives need NVMe. Most reads bound for the data drive stop at the buffer pool, which makes it a hard disk drive candidate. Backups belong on HDD.

The transaction log is the drive that decides how the app feels. Every commit waits on it, so its latency shows up directly in response time. tempdb is the other flash-hungry one.

The data drive is the one people get wrong. SQL Server pulls pages into its buffer pool in RAM and serves repeat reads from there. If the active working set fits in memory, the data drive mostly sees checkpoint flushes and the occasional cold page. It isn’t idle, but it’s nowhere near the monster everyone pictures. And backups? Big sequential streams are the one job spinning disks have always been great at.

Why Most HCI Platforms Say No to Hard Disk Drives

Most HCI platforms treat storage as one cluster-wide pool run by policy, and that’s where the no comes from. VMware’s current direction, vSAN ESA, requires all-NVMe hardware, so HDD isn’t even on the menu. Hybrid designs put a flash cache in front of disk and, by default, let an algorithm decide what stays hot. Some let you pin specific disks to flash, but the starting assumption is that the cache will figure it out. That cache-first design is a big part of why hybrid storage failed the first time.

When the cache guesses wrong, the database eats full HDD latency at a moment nobody picked. So vendors play it safe and spec all-flash for anything that smells like a database. That means paying flash prices for backup files and cold data pages that a spinning disk would serve just fine.

Moving the Data Drive to HDD, Live

VergeOS builds storage into the same code base as compute and networking, and its built-in vSAN sorts media into six tiers. Three of them matter here:

  • Tier 0 holds metadata on enterprise NVMe, so block lookups never wait on a spinning disk, even when the data itself lives on HDD.
  • Tier 1 is performance NVMe, where latency-sensitive drives live.
  • Tiers 4 and 5 are high-capacity hard drives.

Placement is a per-drive setting called Preferred Tier. In the example, the OS, log, and tempdb drives stay on Tier 1 NVMe, and the backup drive already lives on Tier 4. Moving the data drive is one change:

  1. Open the VM and select the data drive.
  2. Change its Preferred Tier from 1 to 4 and save.
  3. That’s it. VergeOS starts a live, online background migration, and the VM keeps serving queries while the blocks move.
Moving a SQL Server data drive from Tier 1 NVMe to Tier 4 hard disk drives, live Animated diagram. The 200 GB data drive moves from the Tier 1 NVMe tier to the Tier 4 hard disk drive tier after one Preferred Tier change while the VM stays online. The VM’s storage media cost drops from $263 to $133, a $130 saving, and 69 percent of its capacity ends up on hard disk drives. SQL Server VM Online, serving queries Storage media cost $263 $133 Saved $130 (49.5%) Moving live, no downtime Move complete 69% of capacity on HDD Tier 1 · NVMe OS, transaction log, tempdb OS80 GB Log40 GB tempdb40 GB Tier 4 · Hard disk drives Backups, cold data Backups150 GB Preferred Tier 1 → 4 Data files 200 GB
One Preferred Tier change moves the 200 GB data drive to Tier 4 hard disk drives while the VM stays online. Media cost drops from $263 to $133.

And placement isn’t a one-time decision. Business needs change, so tiers can too. Quarter-end close coming up? Promote the finance database’s data drive to Tier 1 for the crunch, then drop it back to Tier 4 when it’s over. An app launch, an audit, a seasonal spike: same idea. The move runs live in the background, so storage follows the business instead of the other way around.

I ran exactly this kind of move on a live SQL workload and recorded the whole thing. Here’s the honest part about what to measure. Commit latency is the wrong number to brag about, because the transaction log never left NVMe. The numbers that tell you whether the data drive is happy on HDD are read latency on that drive and PAGEIOLATCH waits in SQL Server. Those are the numbers to watch when you try this on your own database, and What Trips People Up below covers what can push them the wrong way.

The ROI on One Database

Same street prices as our last post: $706 per terabyte for a 15.36 TB NVMe SSD, $56 per terabyte for a 24 TB hard drive. On flash, the 200 GB data drive costs $141.20 in raw media. On HDD, it costs $11.20. That’s 92% less for the same bytes.

SQL Server VM media cost and capacity before and after moving the data drive to hard disk drives Before the move the VM has 360 GB on flash and 150 GB on hard disk drives, for $262.56 in raw media. After moving the 200 GB data drive, it has 160 GB on flash and 350 GB on hard disk drives, for $132.56, a $130 saving or 49.5 percent. Prices are $706 per TB for NVMe flash and $56 per TB for hard disk drives. Flash at $706/TB Hard disk drives at $56/TB Media cost Before $262.56 After $132.56 −$130 (49.5%) Capacity Before 360 GB flash 150 GB HDD After 160 GB flash 350 GB HDD, 69% Raw media cost, single copy. Street prices checked September 28, 2026.
One drive move cuts the VM’s media cost by $130 and puts 69% of its capacity on hard disk drives.

One Preferred Tier change saved $130 and cut the VM’s media cost roughly in half (49.5%), with 69% of its capacity now on hard disk drives.

A $130 saving is small. That’s the point. This is one modest VM. Every terabyte you move from flash to HDD saves $650 in raw media, and a redundancy scheme that keeps a second copy doubles it. Move 20 TB of cold data files and backups across a cluster and that’s $13,000 in raw media, or $26,000 with the second copy counted. VergeOS licenses per server, so that money stays in your budget instead of feeding a capacity meter.

What Trips People Up

  • A working set bigger than RAM. Heavy OLTP that can’t fit its hot data in the buffer pool reads from the data drive constantly. That drive stays on Tier 1, full stop.
  • Maintenance jobs read everything. Index rebuilds, DBCC CHECKDB, and statistics updates scan the data drive end to end. On HDD they take longer, so test your maintenance window, not just your daytime queries.
  • Cold starts. After a restart or failover the buffer pool is empty, and the first wave of queries comes off disk. Expect a slower warm-up on HDD.
  • Checkpoints are light, but random. Random writes are a spinning disk’s weak spot. Watch write latency on the data drive during checkpoints, not just reads.
  • Guessing instead of experimenting. This is the big one. Nobody, me included, can tell you the right tier for your database from a blog post. The move is live and reversible, so treat it as an experiment: move the data drive to Tier 4, watch read latency and wait stats for a week, and move it back if the numbers disagree with you.

The best candidates for HDD: reporting databases, historical and archive databases, dev and test copies, and apps with a small hot set.

HCI Storage Placement Compared

All-flash HCIHybrid cache HCIVergeOS
Role of HDDNoneCapacity behind a flash cacheDedicated tiers (4 and 5)
Who decides where data livesCluster policyCaching algorithm by default; some allow pinningAdministrator, per VM drive
Metadata locationFlashVaries by designTier 0 NVMe
Changing placementNot applicablePolicy changeLive background migration, no downtime
Key Terms

Preferred Tier

A per-drive VergeOS setting that tells the vSAN which storage tier should hold that drive’s data. If the chosen tier doesn’t exist, VergeOS uses the nearest available tier.

Buffer pool

The memory SQL Server uses to cache data pages. It absorbs most repeat reads before they ever reach the data drive.

Live tier migration

Moving a drive’s data to a new tier in the background while the VM keeps running.

What This Adds Up To

A database isn’t a flash workload. It’s a handful of workloads, and each one deserves the level of performance it actually needs, at a price that makes sense. Flash isn’t getting cheaper, so that tradeoff will matter more with every refresh cycle. You’ll have to experiment to find the right balance for your environment, and VergeOS makes that experiment cheap: change a drive’s Preferred Tier, measure, and change it again whenever the business needs something different. Try it Monday, undo it Friday.

George Crump and I did this live: we moved a running VM from flash to hard drives and back, then walked through the refresh math for a real cluster. The full presentation is online now, with every demo included and no registration required. Watch How to Use HCI and Hard Drives to Fight Data Center Inflation.

Frequently Asked Questions

Can SQL Server run on hard disk drives in production?

Yes, for the right drives. Keep the transaction log and tempdb on NVMe. Data files for databases whose working set fits in memory, and backup files, can perform well on HDD. Test before you commit.

How much does moving a database to hard disk drives save?

At September 2026 street prices, each terabyte moved from NVMe to HDD saves about $650 in raw media. In the example above, one 200 GB data drive cut the VM’s media cost from $263 to $133.

Does changing a drive’s Preferred Tier require downtime?

No. VergeOS runs the move as a live, online background migration while the VM keeps running.

How do I pick the right tier for each drive?

Experiment. Start with the drives least sensitive to latency, like backups and cold data files, move them to HDD, and watch latency and wait stats. Because VergeOS moves data live, you can move a drive back if performance doesn’t meet your needs.

Is a drive’s tier placement permanent?

No. A drive can move between tiers at any time, live, as business needs change. Promote it to NVMe for a busy period and move it back to HDD afterward.

Filed Under: Storage Tagged With: , , , , , ,

by

Flash now costs 12 to 14 times more per terabyte than a hard drive, and that gap changes how an HCI cluster gets built.

A 24 TB hard drive sells for about $1,350 today, and a 15.36 TB enterprise NVMe SSD sells for $10,850 or more, a gap wide enough to reopen a storage decision most IT teams considered closed years ago. Dave Vincent and I ran into it this month while pricing drives for our labs, where the SSD quote came in near $12,000 and the hard drive quote barely registered by comparison.

Key Takeaways
  • At September 2026 list prices, flash costs 12 to 14 times more per terabyte than a 24 TB hard drive.
  • All-flash HCI ties every terabyte of capacity growth to the most inflated component in the data center.
  • VergeOS runs hard drives and flash in one cluster, moves running VMs between them on command, and protects the hard drive tier with frequent snapshots and ioGuardian.

The Math Behind a $10,000 SSD

These are street prices, checked on September 28, 2026, that show the gap, and they move week to week right now.

DriveCapacityPriceCost per TB
Kioxia KCMYXRUG15T3 NVMe15.36 TB$10,850$706
Solidigm D7-PS1010 NVMe15.36 TB$12,500$814
WD Ultrastar DC HC580 SATA24 TB$1,350$56

The price of one 15.36 TB SSD buys eight 24 TB hard drives, 192 TB of raw capacity for the same money. VergeIO calls this data center inflation, hardware costs rising faster than IT budgets, driven by AI demand for memory and flash. Flash and memory prices climbed 70 to 95 percent through the first half of 2026, and AI buildouts absorb new fab capacity as fast as it comes online. I plan as if this pricing is permanent.

24 TB hard drive versus 15.36 TB SSD, cost per terabyte Kioxia 15.36 TB NVMe SSD at $706 per TB, Solidigm 15.36 TB NVMe SSD at $814 per TB, and a WD 24 TB hard drive at $56 per TB, based on street prices checked September 28, 2026. Cost per terabyte, street prices, September 28, 2026 $0 $200 $400 $600 $800 Kioxia 15.36 TB NVMe SSD: $10,850, or $706 per TB Kioxia 15.36 TB NVMe $10,850 per drive $706 Solidigm D7-PS1010 15.36 TB NVMe SSD: $12,500, or $814 per TB Solidigm 15.36 TB NVMe $12,500 per drive $814 WD Ultrastar DC HC580 24 TB hard drive: $1,350, or $56 per TB WD 24 TB hard drive $1,350 per drive $56 12 to 14 times less per terabyte
A 24 TB hard drive costs $56 per terabyte against $706 to $814 for a 15.36 TB NVMe SSD. Hover a bar for drive prices.

All-Flash HCI Turns a Media Choice into a Refresh Bill

Live Webinar · October 7 · 2:00 PM ET

How to use HCI and Hard Drives to Fight Data Center Inflation

George Crump and Dave Vincent move a running VM from flash to hard drives and back, live from Palo Alto.

Register Now

Every HCI platform sold today defaults to all-flash, a decision the platform made for the customer years ago, back when flash prices fell every quarter and the assumption looked safe. That default now means every terabyte added at refresh time pays the flash price, and capacity-based licensing charges a second time for the same terabyte.

Most workloads need far less performance than all-flash delivers. File shares, VDI profiles, dev and test environments, log retention and a large share of general-purpose VMs sit idle for most of the day. Hard drives are slower than flash, but they stay online, handle most of this work at a speed users accept and cost a fraction as much.

Aging All-Flash Array? Let the Refresh Replace It

An aging dedicated all-flash array faces the same price sheet at its next refresh. The array vendor buys NAND from the same fabs that supply every SSD in the table above, so the quote for new controllers or more capacity arrives with data center inflation already built in. Many arrays also add a per-terabyte software license on top of the hardware, which charges the inflated flash price a second time. For many IT teams, that refresh falls in the same year as a VMware renewal.

Use the refresh to replace the array and exit VMware in one project. VergeOS runs virtualization, storage and networking on standard x86 servers, so the next purchase buys off-the-shelf flash at server prices instead of flash sold through an array vendor. Global inline deduplication means you buy less of that flash, and the file shares, VM templates, aging VMs and snapshots that sit idle on the array today move to 24 TB hard drives at $56 per TB. The separate storage network and the array support contract go away with it. The Storage Refresh vs. Storage Inflation datasheet walks through the math.

Replacing an aging all-flash array and VMware hosts with VergeOS at refresh Animated diagram. Today, VMware hosts reach a dedicated all-flash array over a separate storage network, and the working set, file shares, VM templates, aging VMs and snapshots all pay the flash price. At refresh, VergeOS replaces the hosts and the array. The working set runs on off-the-shelf flash inside the servers, and everything else moves to 24 TB hard drives in the same cluster. Today At refresh, with VergeOS VMware hosts separate storage network Dedicated all-flash array Working set File shares VM templates Aging VMs and snapshots Every terabyte pays the flash price Replaced at refresh refresh VergeOS cluster ServerServerServer FlashFlashFlash 24 TB24 TB24 TB HDDHDDHDD Less flash, off the shelf. The rest at $56 per TB.
At refresh, VergeOS replaces the VMware hosts and the array. The working set runs on off-the-shelf flash inside the servers, and everything else moves to 24 TB hard drives.

Can a 24 TB Hard Drive Keep Up?

For most workloads, yes. Random-I/O databases still belong on flash, but for the rest of the data center the answer comes down to the platform in front of the drive. One of our customers ran a hosting business that sold both flash and spinning-disk capacity with VergeOS underneath. Their customers tested their workloads on hard drives, saw little difference, and kept moving more of them to disk. Dave Vincent saw the same result when he moved a live SQL workload from NVMe to spinning disk and commit latency stayed under a millisecond on both tiers.

VergeOS makes that result repeatable for any IT team. Global inline deduplication runs across every tier, so the cluster stores fewer blocks and more of the working set fits in cache. VergeOS caches data in RAM on the server running the VM, which VergeIO testing shows produces four to five times the cache hit rate of array-side caching. On the write side, VergeOS reorders and groups incoming writes into sequential runs before they reach the platter, the access pattern hard drives handle best.

How VergeOS makes a 24 TB hard drive tier keep up Animated diagram of the VergeOS I/O path. Reads are served from a RAM cache on the server running the VM. Random writes pass through global inline deduplication and are reordered and grouped into sequential runs before they reach the hard drive. VM any workload RAM cache on the server running the VM Reads served from RAM, 4 to 5 times the hit rate of array-side cache Random writes in Dedupe, reorder and group fewer blocks, longer runs Sequential runs out 24 TB hard drive Hard drives handle long sequential runs best. VergeOS shapes the write stream to match before it reaches the platter.
Reads come from RAM next to the VM. Writes arrive random and leave as sequential runs, the pattern a spinning drive handles best.

Afraid of Tiering? Select Tiering That Puts You in Control

Many IT teams steer clear of tiering, and automated tiering is the reason, after two decades of burning plenty of them. An auto-tiering algorithm samples activity, waits for a threshold and makes placement calls without knowing the business behind the workload. It is slow to spot cold data, it misses anything outside its sampling window, and it is slowest of all on the trip back to flash, usually right when users start to notice.

VergeOS puts the IT team in charge of placement instead. The people who run the workloads already know the business calendar, such as a tax application that is write-heavy through April 15 and read-heavy on April 16, and they need no algorithm to predict it. The IT team moves a running VM to the hard drive tier or back to flash on command, in either direction and with no maintenance window, and pins the workloads that must stay hot to a preferred tier. I covered why tiering itself was never the issue in Storage Tiering Is Not the Problem. Losing Control of It Is.

VergeOS moves a running VM between the flash tier and the hard drive tier on command Animated diagram. A running VM sits on the Tier 1 NVMe flash tier, moves down to the Tier 4 hard drive tier on command with the workload online, then moves back to flash the same way. Tier 1 NVMe flash Working set, random-I/O databases Tier 4 24 TB hard drives File shares, templates, aging VMs Move down on command Move back Running VM online the whole time Lab test: API acknowledged the move in 1.3 seconds. A 25 GB disk finished moving in 13 seconds.
The IT team decides placement. A running VM moves to the 24 TB hard drive tier and back on command, with no maintenance window.

Three Ways VergeOS Makes Hard Drives Safe to Use

Tiering under IT control. VergeOS supports six storage tiers inside one vSAN, with tiers 4 and 5 designed for high-capacity hard drives, and the IT team moves a running VM’s disk between tiers on command. In VergeIO lab testing, the API acknowledged the move in 1.3 seconds and a 25 GB disk finished moving in 13 seconds with the workload running. Be aggressive with it, moving a workload to hard drives and moving it back in seconds if someone complains two weeks later.

Frequent snapshots. VergeOS snapshots run on a schedule the IT team controls, hourly by default, which gives every VM on the hard drive tier recent recovery points inside the same system.

ioGuardian. Large drives raise a fair question about exposure during a rebuild. ioGuardian answers it by pulling missing or corrupted blocks from a synchronized remote VergeOS system after a failure that exceeds the cluster’s redundancy, such as multiple drive failures spread across nodes.

VergeOS licenses per server, so the money saved on hard drives and deduplication stays in the budget instead of feeding a capacity meter.

Key Terms

Data center inflation

Hardware costs rising faster than IT budgets, driven by AI demand for memory and flash.

Storage tier

A class of drives in the VergeOS vSAN, numbered 0 through 5, running from metadata NVMe to archival hard drives.

Preferred tier

The VergeOS setting that names the first-choice tier for a VM disk, used to keep critical workloads on flash.

ioGuardian

The VergeOS repair server that restores blocks from a synchronized remote system after a failure beyond redundancy tolerance.

All-Flash HCI and VergeOS Flash Plus Hard Drives

All-flash HCIVergeOS flash plus hard drives
Capacity cost at refreshFlash price for every TBFlash for the working set, $56 per TB for the rest
Workload placementEverything on flashIT team places each VM, moves it live
LicensingOften per TBPer server
Media mixFlash onlyFlash and hard drives in one cluster

See the Answer Live on October 7

Dave and I will answer the question in the title live from Palo Alto on Wednesday, October 7, at 2:00 PM ET, in How to use HCI and Hard Drives to Fight Data Center Inflation. Dave will move a running VM from flash to hard drives and back on camera, and I will walk through the refresh math for a real cluster, so register for the webinar.

Frequently Asked Questions

Can a 24 TB hard drive replace an SSD for every workload?

Random-I/O databases still belong on flash, and VergeOS lets you pin them to a preferred flash tier and place everything else on hard drives.

Does deduplication work on the hard drive tier?

VergeOS runs global inline deduplication across all six tiers, so the hard drive tier gets the same capacity savings as flash.

How long does it take to move a VM back to flash?

The move starts on command and runs with the VM online, and in lab testing small disks finish the trip back to flash in seconds.

Filed Under: Storage Tagged With: , , , , ,