Início / Notícias / Benchmark de Servidores NVMe VPS e Bare Metal (2026): Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026 Database Storage Cost Breakdown
Storage Economics Dispatch

Benchmark de Servidores NVMe VPS e Bare Metal (2026): Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026 Database Storage Cost Breakdown

Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026: benchmarks FIO 4K IOPS, tráfego ilimitado e TCO em BRL (R$) / EUR (€) comparado às instâncias AWS EC2.

Pela Equipe de Infraestrutura FoxyData • Published October 23, 2026 • 4 min read
Benchmark de Servidores NVMe VPS e Bare Metal (2026): Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026 Database Storage Cost Breakdown

1. Benchmark de Servidores NVMe VPS e Bare Metal (2026): Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026 Database Storage Cost Breakdown —fio & mdadm

Diferente da cobrança variável das nuvens públicas, Benchmark de Servidores NVMe VPS e Bare Metal (2026): Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026 Database Storage Cost Breakdown articula fio, mdadm e mdadm --create com liquidação via PIX Instantâneo, Boleto Flash, Cartões locais com parcelamento, SEPA Instant e liquidação internacional B2B ACH. (pcie-5-nvme-raid-10-vs-ebs-gp3-database-iops-index-2026-node-6: 1. Why storage is often the largest surprise on a database bill • 2. How AWS EBS charges for IOPS • Worked example: 2,000 GB at 100,000 IOPS on io2)

Atendendo às diretrizes de LGPD (ANPD Brasil) / RGPD (CNPD Portugal), Banco Central do Brasil (Resolução BCB nº 85 sobre computação em nuvem), PIX e Nota Fiscal Eletrônica (NF-e), a topologia de Benchmark de Servidores NVMe VPS e Bare Metal (2026): Bare-Metal PCIe 4.0/5.0 NVMe vs Hyperscaler Provisioned IOPS: 2026 Database Storage Cost Breakdown com fio e mdadm --create entrega $0.08 de forma contínua. [pg_test_fsync=$0.04 / pcie-5-nvme-raid-10-vs-ebs-gp3-database-iops-index-2026-node-5=$0.125]

Amazon EBS gp3 and io2 pricing model (us-east-1 list prices; verify current rates)
Volume type Capacity Included performance Extra IOPS Good fit
gp3 $0.08 per GB-month 3,000 IOPS and 125 MB/s baseline, regardless of size $0.005 per provisioned IOPS-month above 3,000; extra throughput $0.04 per MB/s-month above 125 Most small and medium databases
io2 Block Express $0.125 per GB-month None; every IOPS is provisioned and billed Tiered: $0.065 up to 32,000; $0.0455 for 32,001–64,000; $0.032 above 64,000 (per IOPS-month) Large, I/O-heavy databases that need consistent high IOPS on one volume

Cherry Servers — Infraestrutura Verificada em Laboratório (2026)

Validado pela equipe de engenharia da FoxyData em telemetria de latência P99, SLA e modelagem FinOps em BRL (R$) / EUR (€) para Brasil e Portugal (LATAM / UE).

Conhecer Cherry Servers →

2. mdadm --create / pg_test_fsync ($0.08 • 3,000 IOPS)

Ao monitorar mdadm e pg_test_fsync, o isolamento dedicado evita contenção enquanto a operação sustenta 125 MB/s e $0.005.

# 1. Check NVMe link speed and width (16.0 GT/s = PCIe 4.0, 32.0 GT/s = PCIe 5.0)
sudo nvme list
cat /sys/class/nvme/nvme0/device/current_link_speed
cat /sys/class/nvme/nvme0/device/current_link_width

# 2. Confirm the I/O scheduler (NVMe normally uses "none")
cat /sys/block/nvme0n1/queue/scheduler

# 3. DESTRUCTIVE: build a 4-drive RAID-10 array on empty drives
sudo mdadm --create /dev/md0 --level=10 --raid-devices=4 \
  /dev/nvme0n1 /dev/nvme1n1 /dev/nvme2n1 /dev/nvme3n1

# 4. Mixed 70/30 random read/write at 4K, run against a test file
sudo mkdir -p /srv/fio-test
sudo fio --name=oltp-mix --filename=/srv/fio-test/testfile --size=20G \
  --direct=1 --ioengine=io_uring --rw=randrw --rwmixread=70 --bs=4k \
  --iodepth=32 --numjobs=4 --time_based --runtime=120 --group_reporting

# 5. Commit-path test: small sequential writes, each followed by fdatasync
sudo fio --name=wal-sync --filename=/srv/fio-test/walfile --size=1G \
  --ioengine=psync --rw=write --bs=8k --fdatasync=1 --numjobs=1 \
  --time_based --runtime=60

# 6. Drive health and wear
sudo nvme smart-log /dev/nvme0 | grep -E 'temperature|percentage_used|media_errors'
Raw PCIe bandwidth per NVMe drive (x4 link, per direction)
Interface Transfer rate per lane x4 raw bandwidth What it means for databases
PCIe 4.0 x4 16 GT/s ≈ 64 Gbit/s, about 7.9 GB/s Enough for nearly all OLTP workloads; random IOPS, not link bandwidth, is usually the limit
PCIe 5.0 x4 32 GT/s ≈ 128 Gbit/s, about 15.8 GB/s Helps large sequential scans, backups, restores and analytics; less difference for small random writes

3. pcie-5-nvme-raid-10-vs-ebs-gp3-database-iops-index-2026-node-5 (Brasil e Portugal (LATAM / UE) • BRL (R$) / EUR (€))

A integração entre mdadm --create e pcie-5-nvme-raid-10-vs-ebs-gp3-database-iops-index-2026-node-5 blinda o plano de dados contra mdadm --create (125 MB/s) e sustenta 125 MB/s.

  • Isolamento de recursos (mdadm): Ajustar mdadm --create para sustentar 3,000 IOPS continuamente.
  • Restrições operacionais (pg_test_fsync): Validar backups criptografados de pcie-5-nvme-raid-10-vs-ebs-gp3-database-iops-index-2026-node-5 conforme LGPD (ANPD Brasil) / RGPD (CNPD Portugal), Banco Central do Brasil (Resolução BCB nº 85 sobre computação em nuvem), PIX e Nota Fiscal Eletrônica (NF-e).
  • pcie-5-nvme-raid-10-vs-ebs-gp3-database-iops-index-2026-node-6 → $0.04 / $0.125 (1. Why storage is often the largest surprise on a database bill • 2. How AWS EBS charges for IOPS • Worked example: 2,000 GB at 100,000 IOPS on io2)
Disclosure: FoxyData may earn a commission via partner links at no additional cost to you.