Octopus Mode is an architectural framework designed to minimize transactions landing time on the Solana blockchain. It was born out of a need to overcome technical bottlenecks, achieve extremely low latency, and provide scalability to users operating in highly competitive environments.

Octopus Mode distributes transaction load across dozens of globally distributed servers located where the most relevant stake is concentrated. Its core idea is to “extend the octopus’s tentacles” wherever they’re needed, strategically placing nodes in cities like Frankfurt, New York City, Amsterdam, Vilnius, Ashburn, Tokyo, and more, to ensure ultra-low-latency connections to Solana leaders and maximize performance.

Think of it like Cloudflare, built on Solana.

What problems does Octopus Mode solve?

• Overcoming network limitations: Direct connections to the TPU (Transaction Processing Unit) of Solana leaders suffer from jitter and aren’t always reliable. Octopus Mode shifts latency to the P9 Nodes fully controlled internal network, leveraging trusted direct connectivity for transaction delivery.
• Stake usage efficiency: Nowadays nodes need a large amount of SWQOS (Stake-Weight-Quality-Of-Network) to land transactions, and they overcome this issue by usually partnering with external validators in the multiple regions they offer (at least one validator per region is needed). With Octopus Mode, stake usage efficiency is significantly boosted, as the whole stake is used in every single location at the same time. This means that it doesn’t matter from where you send your transaction, you will always connect to all the stake under P9 Nodes’ control. This radical change offers the possibility of future cluster expansion, without the need of more resources.
• Anti-MEV blacklist protections [beta]: Octopus Mode defends users against sandwich-labeled validators by dynamically skipping them entirely (Protection On), or by sending transactions to them only in limited scenarios (Protection Off). Users are able to choose which protection works the best for them by just changing port on their endpoints. We reduced drastically the statistical probability of being sandwiched with both endpoints, with best results on “Protection On” endpoint. [This feature is in beta and will be updated day by day to have the best ratio between skipped blocks and % of sandwich rate]

How does it work?

• Distributed architecture: The cluster combines high-performance full nodes, geyser nodes, and a distributed “spiderweb” of nodes running our custom optimized SendTx engine, all tied together by load balancing solutions that allows the infrastructure to flexibly scale with user demand.
• Staked and Unstaked connections: Octopus Mode takes advantages of both “staked” (high-reliability, high-capacity) and “unstaked” (numerous, lower-reliability) endpoints. Transactions are sent in parallel across both classes to ensure the best landing time.

Whats coming alongside Octopus Mode