Solana's engineering north star is IBRL—Increase Bandwidth, Reduce Latency. Every major upgrade shipping today has that goal in mind. As the network gets faster, the gap between the protocol and the hardware it runs on keeps shrinking.
The closer the protocol gets to the hardware, the less room there is for the abstraction layers that cloud providers and container platforms put in between. Solana validators can run on AWS, GCP, or inside a container – but in practice, validators running on bare metal hardware perform better under load.
100M CUs
As the ecosystem prepares to increase from 60M to 100M compute units (CUs) per block (a 66% increase), the need for bare metal hardware becomes critical. At higher CUs, the primary constraint becomes Turbine, the layer that propagates blocks across the network. If shreds can't fan out to thousands of nodes fast enough, the extra capacity doesn’t translate to meaningful improvements to the network.
XDP: High Performance Networking
XDP (eXpress Data Path) is high-performance mode for your network interface card. It skips the slower more generalized path that your kernel uses to handle networking and instead moves the logic to the hardware. To support the 100M CU feature activation, XDP will soon be enabled by default for all validator clients.
The Anza XDP setup guide walks operators through XDP configuration and explains what operators must do to get the most performance from their hardware.
Here are a few requirements to highlight:
-
Elevated capabilities The validator process needs
CAP_NET_RAW,CAP_NET_ADMIN,CAP_BPF, andCAP_PERFMON -
Dedicated cores XDP and Proof of History (PoH) must be assigned to separate physical cores. Not threads, not "vCPUs"—physical cores.
-
Serious packet rates Because Turbine fans shreds out aggressively, a highly staked validator can push approaching 150,000 outbound packets per second. Highly staked nodes send even more, because they get more leader slots.
These requirements are easy for validators to satisfy when they have control over the hardware, since validators can pick the NIC, pin the cores, and choose the driver and kernel.
Why Bare Metal Performs Better
Bare metal gives the validator direct access to the hardware. With no virtualization in between, XDP talks to the NIC directly, which is where it performs best.
On bare metal, you own the NIC. You pick the model, firmware, driver, and queue implementation, and you control the switch path. As network demand ramps up, operators should prefer the most performant hardware for the network card.
The hardware choice and direct access means predictable performance under load. When Turbine fans shreds out at 150,000 packets per second during your leader slots, the full NIC and CPU are available to you. The same goes for storage: you choose known enterprise NVMe drives instead of a network-attached block device with provisioned IOPS limits.
The community has created a very useful resource, Solana Hardware Compatibility List, cataloging common Solana validator hardware along with a summary of operator opinions on the hardware. The site contains known good CPUs, storage, and networking for mainnet validators, and recommends dedicated hardware. A good example from the site is Anza's networking requirement for a staked node is a 2 Gbit/s symmetric connection, but the community recommendation of 10–25GbE comes from experience running a high performance machine in practice.
Containers add a layer you then have to delete
Containers add another layer of abstraction that can lead to performance
issues. The only container shape that preserves XDP performance is one that
systematically removes container isolation: --network=host to share the
host's network namespace and the elevated capabilities listed above allow for
direct access to the host interface. This direct hardware access makes the
container unnecessary.
Anza's own validator requirements strongly suggest that running an Agave validator for live clusters, including mainnet, inside Docker is "not recommended and generally not supported," citing containerization overhead and performance degradation unless specially configured. The same page warns that running in the cloud "requires significantly greater operational expertise to achieve stability and performance."
Summary
In order to run a high-performance Solana validator, it is strongly recommended to run on dedicated bare metal. Prefer a high-clock CPU, ECC RAM, fast enterprise NVMe drives, and 10 to 25GbE symmetric connectivity. Given the upcoming 100M CU block capacity, operators should choose high performance modern NICs and run the validator directly, without containers.
References: