Scaling the Read Layer to 1M TPS: FluxBeam / FluxRPC / RugCheck

As Solana prepares for Firedancer's promised million transactions per second, FluxBeam has unveiled a groundbreaking solution to what could become the network's next major bottleneck: reading all that data. At Breakpoint 2025, Scott Hague, technical co-founder of FluxBeam, introduced Lantern—a revolutionary RPC architecture delivering a staggering 99% reduction in latency and up to 90% cost savings for developers.

Summary

Top Projects

Lulo

Score: 7,586.13

Bonk

Score: 4,806.12

Sphere

Score: 2,544.42

Moonwalk

Score: 2,026.11

The Solana ecosystem faces a critical challenge that many haven't fully appreciated yet. While Firedancer promises to dramatically increase the network's transaction throughput to potentially one million TPS, the existing infrastructure for reading blockchain data wasn't designed for this scale. FluxBeam recognized this problem early and spent the past year rebuilding the entire RPC layer from scratch.

Traditional RPC setups rely on stacking non-voting validators with RPC layers on top, routing client requests to individual servers. This approach has already shown its limitations—anyone who has tried to call get program accounts on Pump Fun has likely experienced timeouts or silent failures due to overwhelming data loads. Multiple clients competing for resources on the same validator lead to lagging slots, stale data, and the leader tracking issues that have plagued latency-sensitive applications like RugCheck.

FluxBeam's solution decouples the read layer entirely from the consensus layer, enabling independent scaling on commodity hardware rather than expensive specialized servers. The centerpiece of their innovation is Lantern, a local RPC that developers deploy directly in their data centers, eliminating network latency entirely. Combined with their DeltaStream technology—which transmits only the bytes that change rather than full account updates—the system reduces bandwidth requirements from 10 gigabits to just 100 megabytes.

Key Points

The Problem: RPC Infrastructure Can't Keep Up

The legacy approach to Solana RPC infrastructure has served the ecosystem reasonably well for several years, but cracks are appearing as network usage intensifies. The traditional model involves deploying multiple non-voting validators, each with an RPC layer stacked on top, with clients routed to individual servers. This architecture creates several critical problems that will only worsen as throughput increases.

When multiple clients compete for resources on the same validator, the system struggles to maintain performance. Data becomes stale, slots lag behind, and leader tracking—essential for time-sensitive applications—becomes unreliable. For teams building applications where milliseconds matter, like the RugCheck token analysis platform, these limitations have become increasingly painful. The anticipated arrival of Firedancer, with its million-TPS capabilities, would transform these issues from inconveniences into showstoppers.

Decoupled Architecture: A Microservices Approach

FluxBeam's fundamental innovation involves completely separating the read layer from the consensus layer. This architectural decision enables the read infrastructure to scale independently, much like microservices in modern web development. Rather than requiring expensive, monolithic hardware setups, the new system runs effectively on commodity hardware and works seamlessly in cloud environments.

The microservices approach also delivers significant redundancy benefits. Each layer can fail independently without bringing down the entire system, and capacity can be added precisely where needed. FluxBeam has also optimized bandwidth utilization, addressing the often-overlooked problem of massive egress bills that come with streaming blockchain data at scale.

Lantern: Your Local RPC

Perhaps the most exciting component of FluxBeam's announcement is Lantern, their local RPC solution tailored specifically for decentralized applications. In traditional web development, applications keep their databases and data stores physically close to their servers. Web3 has historically forced a different, more challenging approach—either constantly querying data over the network or attempting to manage complex ingestion infrastructure internally.

Lantern changes this paradigm entirely. Instead of ingesting the entire Solana data firehose, developers can specify exactly which data their application needs. This filtered data streams in real-time directly into the developer's data center, right alongside their application. Calling data locally eliminates network latency completely, resulting in dramatically improved user experiences. The system features intelligent subscriptions that automatically detect when needed data isn't cached, setting up the necessary pipelines without manual intervention.

DeltaStream: Bandwidth Revolution

A key enabler of Lantern's efficiency is DeltaStream technology, which fundamentally reimagines how blockchain data is transmitted. Rather than sending complete account updates every time something changes, DeltaStream transmits only the bytes that actually changed. This compression technique reduces bandwidth requirements by orders of magnitude—from requiring a 10-gigabit connection to functioning on a 100-megabit pipe.

The bandwidth savings translate directly into cost reductions and scalability improvements. Teams can add more Lantern instances to handle additional RPS (requests per second) without proportional increases in infrastructure costs. The technology makes running sophisticated on-chain data pipelines feasible for teams that couldn't previously afford the bandwidth and infrastructure requirements.

Facts + Figures

• FluxBeam's architecture delivers approximately 20% cost reduction compared to traditional public RPCs
• For private RPC deployments, cost savings reach 80-90%
• Lantern achieves a 99% reduction in latency by eliminating network round trips
• The system requires 99% fewer resources than traditional RPC infrastructure
• DeltaStream reduces bandwidth requirements from 10 gigabits to 100 megabits—a 100x reduction
• The architecture is built specifically to handle Firedancer's projected 1 million TPS
• FluxBeam runs custom Firedancer tiles for data ingestion
• The system runs on commodity hardware rather than specialized expensive servers
• Lantern functions as a zero-touch setup with intelligent automatic subscriptions
• The solution works effectively in cloud environments with optimized egress costs

Top Quotes

• "After seeing the demos from Firedancer last year, we realized there's going to be a big issue coming up around we can do the 1 million TPS from the ingestion side. But how do we then read all of that data coming through?"

• "I think a lot of you have probably called get program accounts on Pump Fun and been met with pretty much silence because of the data load."

• "In Web 3, we like to do things the difficult way."

• "Zero latency means better UX for everyone."

• "Instead of sending you all the account updates of all the pieces of data, we're only sending you the bytes that change. So rather than needing a 10 gig pipe, you can do this on a 100 megabyte pipe."

• "No longer do you need to pay a couple of grand a month for these RPC servers. You can run Lantern locally."

Questions Answered

What is the main problem FluxBeam is solving?

FluxBeam is addressing the upcoming challenge of reading blockchain data at scale as Solana prepares for significantly higher throughput with Firedancer. While the network's transaction ingestion capabilities are advancing to potentially one million TPS, the traditional RPC infrastructure for reading that data wasn't designed to handle such volume. Current systems cause lagging slots, stale data, and timeout issues even at today's transaction levels, problems that would become critical at higher throughput.

How does Lantern differ from traditional RPC setups?

Lantern is a local RPC that runs directly in a developer's data center rather than requiring network calls to remote servers. Unlike traditional setups where applications must query data over the network every time they need information, Lantern streams only the specific data a particular application needs directly to where it's being used. This eliminates network latency entirely and allows developers to query blockchain data with the same speed as local database calls.

What is DeltaStream and why does it matter?

DeltaStream is FluxBeam's technology for efficiently transmitting blockchain data updates. Instead of sending complete account information every time something changes, it transmits only the specific bytes that were modified. This approach reduces bandwidth requirements by approximately 100x—from 10 gigabits to just 100 megabits—making sophisticated blockchain data infrastructure feasible for teams that couldn't previously afford the bandwidth costs.

How much can developers save with FluxBeam's infrastructure?

The cost savings vary depending on the deployment model. For public RPC usage, FluxBeam's architecture offers approximately 20% cost reduction compared to traditional providers. For teams running private RPC infrastructure, the savings are far more dramatic—reaching 80-90% reduction. Additionally, the 99% decrease in resource requirements means developers can run effective blockchain data infrastructure on much more affordable hardware.

Can the system scale with my application's needs?

Yes, the architecture is specifically designed for horizontal scaling. Because the read layer is completely decoupled from the consensus layer, developers can add more Lantern instances to handle increased requests per second without complex infrastructure changes. The intelligent subscription system automatically handles data routing and caching, making scaling a largely automatic process rather than requiring manual infrastructure management.

Why should developers care about this before Firedancer launches?

Even without Firedancer's million-TPS throughput, current RPC infrastructure already shows strain with popular applications. The problems of stale data, timeout errors, and leader tracking issues affect latency-sensitive applications today. Adopting improved infrastructure now provides immediate performance benefits while also future-proofing applications for the significant throughput increases expected when Firedancer goes live.