SQD Network Explained by DK

We’re kicking off our new “Explained by…” series - short, engineer-written breakdowns of the SQD stack.

First up, DK (core SQD dev) opens the hood on how our network upgrade actually serves data: Portal, workers, hotblocks - no fluff, just how it runs.

What's SQD Network?

SQD Network is a huge and growing compute cluster consisting of 2500+ decentralized workers.

Each worker has 1 TB of storage and is equipped with a powerful query engine, allowing them to serve queries to the Network.

Decentralization means that the workers are not controlled by a single party but are supported by independent operators across the globe.

What's Portal?

Portal, formerly known as Gateway, is a service opening a window to the Network for the user.

It communicates with decentralized workers via libp2p, providing a simple and efficient HTTP API for streaming data.

You can think about it as a bridge connecting centralized and decentralized worlds.

What are Hotblocks?

While static archival data is served from the Network allowing very high throughput, the “last-mile” real-time data requires low latency.

That's why this data is stored together with the Portal and receives live updates from RPC nodes.

The Portal then combines both data sources to provide a single stream of data that has both high throughput and low latency.

Can I run my own Portal?

Yes, and to access SQD Network, you have to lock SQD tokens. The more SQD you lock, the higher rate limit you get.

For most people it's easier to use one of the existing Portals with flexible payment options provided by its operators.

For example, we're providing a public Portal for free to test your code before going to production.

What data is available?

Currently, SQD Network supports raw blockchain data for 200+ networks.

It means that you can request which fields you need, set filters, and get results from the entire blockchain history downloaded within minutes or hours (depending on the amount of extracted data) as opposed to days or months, as it would be with direct RPC requests.

The aggregation (calculating token balances, prices, etc) is then done on the client. We also provide a convenient Software Development Kit (SDK) for that.

What exactly are we releasing at the end of 2025?

We've already been running a centralized version of the Network called "v2 archives".

Switching to the Network gives multiple benefits:

• Hundreds of times bigger compute cluster — Petabytes of storage, thousands of CPUs
• 10x faster Rust query engine
• Much faster data streaming enabled by fetch-ahead parallelism
• "Hotblocks" included - no need to bring your own RPCs anymore
• Permissionless access
• Much bigger datasets, like the full history of Solana and Eclipse

If this helped, follow the series - each post demystifies one layer of SQD, straight from the engineers and executive team. Bring questions; we’ll use them to shape the next brief. Deeper.