SLV Implements Speed Comparison Benchmark Tool for Solana RPC / Geyser gRPC / Shredstream — Visualizing Global Distributed Environment Bottlenecks with Region-Specific Filtering

ELSOUL LABO B.V. (Headquarters: Amsterdam, Netherlands; CEO: Fumitake Kawasaki) and Validators DAO have implemented a benchmark tool in their jointly developed and operated open-source Solana development tool, SLV, to compare the speeds of Solana RPC, Solana Geyser gRPC, and Solana Shredstream.

The standout feature of this tool is its region-specific filtering. While conventional measurement tools compare overall results collectively, SLV's benchmark tool outputs a comparison limited to results from validator leaders in specific regions, in addition to the results from all regions (global). It can be launched with a single command and can be run either interactively or with specified options.

For developers and traders pursuing ultra-low latency, the ability to more accurately grasp the true performance in the region where their servers are located dramatically changes the precision of environment selection and architectural design.

SLV Official Website: https://slv.dev/en Speed Comparison Documentation: https://slv.dev/en/doc/geyser-grpc/speed-comparison/

Why is region-specific measurement necessary? — The structural challenge of a globally distributed blockchain

In Solana, block production involves a validator leader that moves around the world in short cycles. In one slot, a validator in Frankfurt might be the leader; in the next, Tokyo; and then New York — the geographical location of the leader is constantly changing.

Conventional measurement tools aggregate data from all slots at once. While this is a correct approach for general-purpose evaluation, it has structural limitations for those pursuing ultra-low latency.

When a server located in Frankfurt receives data generated by a validator leader in Tokyo, propagation delay due to physical distance is unavoidable. This delay is a geographical handicap that occurs regardless of server or network quality. When measuring all slots collectively, this geographical handicap gets mixed into the numbers, burying the true performance value of your own region in noise.

SLV's benchmark tool solves this problem. By outputting the global results from all regions and the results focused on validator leaders in a specific region separately, it enables an evaluation of infrastructure performance that takes geographical conditions into account.

Two perspectives that change strategy

The results from all regions indicate the overall response quality to any leader. This metric is important for dApps with a global user base or use cases that require stable responses regardless of leader location.

The results from a specific region show the true performance against leaders where your server has a geographical advantage. In use cases like high-frequency trading or arbitrage, where a millisecond advantage can directly impact results, how quickly you can reach the leader in your own region determines victory or defeat.

Which metric to prioritize depends on the use case and strategy. The important thing is that it's now possible to evaluate data separately, which was previously only visible in aggregate. For example, when changing providers, you can check not just the overall win rate, but also whether performance has improved or worsened for leaders in a specific region. The accuracy of environment selection decisions is significantly improved.

Measuring all three communication layers — Solana RPC / Geyser gRPC / Shredstream

SLV's benchmark tool measures all three major communication layers of the Solana infrastructure.

Solana RPC is the interface through which all Solana applications access the network. Solana Geyser gRPC is a real-time stream of on-chain data, used for fetching price feeds and monitoring account states. Solana Shredstream is the layer that receives block data fragments directly from the leader, and is the lowest-latency data acquisition method.

For each of these three layers, a comparison between all regions and specific regions is possible. You can get a cross-sectional understanding of which layer has a bottleneck and in which region the performance difference occurs, all with a single tool.

A measurement foundation that comes alive when combined with SLV's AI agents

The benchmark tool is useful on its own, but when combined with the SLV AI agent environment, it connects the dots from measurement to improvement.

SLV is an AI Agent Kit that covers the entire lifecycle of Solana development. It has specialized agents for each purpose: Cecil for validator operations, Tina for managing indexed Solana RPC nodes, Cloud for configuring Solana Geyser gRPC streaming, and Setzer for assisting with Solana app development, each with specialized knowledge in their respective domains. When a request is made in natural language from the AI Console, it is automatically routed to the appropriate specialized agent based on the content.

The benchmark tool supports the Model Context Protocol (MCP), allowing AI agents to execute measurements and read the results. For example, if a performance difference in a specific region is identified from the measurement results, you can proceed with server procurement, node migration, and configuration changes within the same SLV environment. Server procurement is supported up to the purchase of SLV Metal, and zero-downtime migration is implemented for node migration. Since the measurement tool, migration tool, and tuning tool are on the same MCP foundation, the AI agent can combine these to assist with everything from bottleneck identification to environmental improvement.

Previously, operating a Solana node required deep knowledge of Linux, proficiency with the CLI, and judgment in version management. SLV's AI agents take on the majority of this cognitive load. The addition of the benchmark tool means that AI agents can now also handle the starting point of the operational cycle, "current state measurement," creating an environment where the entire cycle of measurement → judgment → execution → re-measurement can be performed in natural language.

Also supports local mode — SSH into a node and measure directly

In addition to remote management, SLV also supports a local mode. You can log in to a node via ssh and run SLV directly on that node.

The benchmark tool also works as-is in local mode. You can run measurements directly on an operational node, check the results, and, if necessary, make improvements while interacting with the AI agent. Since there's no need to go through a remote management node, the configuration is simple, and you can quickly gauge the performance of the node at hand.

An environment built in local mode can be gradually migrated to a remote management configuration using Ansible as the project grows. Start with one, scale as needed — SLV's design philosophy is consistent, even with the benchmark tool.

Providing crucial data for high-frequency trading decisions

Solana is a globally distributed blockchain. Its structure, where block production is performed sequentially by validators around the world, requires more complex decisions from high-frequency traders and bots than just simple latency.

In which region's leader time is my environment advantageous? In which layer is the difference appearing? If I switch providers, how will my win rate in a specific region change? — SLV's benchmark tool provides the latency comparison, region-specific comparison, and layer-specific comparison needed for these decisions, all in one open-source tool.

Publishing accumulated operational knowledge as open source

SLV's benchmark tool, like its other features, is provided as open source.

SLV is an open-source tool that allows you to perform everything from launching a Solana validator or RPC node to daily operations, primarily through natural language interaction with an AI agent, in a no-code manner. It covers the entire lifecycle of Solana development, including deployment, upgrades, downgrades, zero-downtime migration, server procurement, Solana Geyser gRPC configuration, and Solana app development. This new benchmark tool adds the starting point of "current state measurement" to this operational foundation.

The various features of SLV, including this benchmark tool, reflect the knowledge gained from the practical operation of the ERPC platform. Optimization methods and tuning parameters accumulated through the operation of the Epics DAO validator, which reached 3rd place among all Solana validators globally, and feedback from users in over 100 countries worldwide, are being released as open source through SLV.

Anyone who wants to evaluate and compare Solana operational environments can use it for free.

SLV GitHub: https://github.com/validatorsDAO/slv

WBSO approved for 5 consecutive years

ELSOUL LABO has been approved by the Dutch government's R&D support program, WBSO, for five consecutive years since 2022. The company continues its research and development on Solana RPC infrastructure based on ultra-low latency, and the automation of validator deployment and operational orchestration, the results of which are directly implemented as tools in SLV.