Ultimate Solana glossary: 2026 edition
TL;DR
- Solana uses Proof of History (PoH) as a shared, cryptographically verifiable clock to *pre-order* events before consensus. PoH itself is not the full consensus mechanism. Solana uses PoH alongside a Proof-of-Stake BFT consensus (called Tower BFT) to agree on blocks and finalise state.
- The state is separated from logic. Unlike the EVM, where contracts store their own state, Solana separates the executable (Program) from the storage (Account).
- Parallelism is the default. The Sealevel runtime executes non‑overlapping transactions in parallel, like a multi‑core CPU running independent threads.
- We’ve put together a glossary of the terms you’ll hear when talking about Solana. From this article, for each core term you’ll get:
- a definition (what it is)
- a function on Solana (what it does)
- a simple mental model / you can use to understand its function and definition better
Terms marked with 🟢 may change or evolve with upcoming protocol upgrades such as Alpenglow.
1. Time & consensus
TL;DR
Most blockchains spend more effort arguing about when something happened than actually processing it. Solana flips this by establishing a cryptographically verifiable clock before consensus even starts. Note: Solana still relies on a Proof-of-Stake consensus algorithm (Tower BFT) to finalise the agreed history. PoH provides ordering that makes consensus more efficient, but does not replace it.
Imagine an endless ship’s logbook (ledger).
- The leader is the captain on shift.
- A slot is a single page in that logbook.
- The leader's schedule is the roster that tells us who the captain is for each page.
- If a captain is “sick” (down), that page number still exists but stays blank (skipped slot), and the book continues.
2. Memory & state
TL;DR
Crucially, Solana separates logic from data. The logic (program) does not store user balances or NFTs within itself; it processes data stored in a separate state (accounts). This decoupling is a key design choice that enables Solana’s performance and parallelism. To understand it, imagine Solana as a global computer, where
- Accounts are files.
- Programs are applications that read/write in those files.
- Rent is the cost of occupying storage space.
3. Execution & runtime
TL;DR
Imagine that inside every Solana validator, there’s a commercial kitchen optimised for throughput.
- Sealevel is a kitchen layout that lets multiple chefs (CPU cores) cook simultaneously.
- Accounts are the ingredients each chef needs.
- Transactions are order tickets.
- Instructions are individual recipe steps.
If Chef A is cooking steak and Chef B is making salad, they can cook in parallel. They only wait if they both need the same ingredient (write‑lock contention on the same account).
4. Network & infrastructure
TL;DR
We’ve talked about time and execution. Now we need to move data and blocks between machines.
- A cluster is a specific network (e.g. Solana mainnet‑beta) and all the elements inside of it.
- Nodes are servers within the cluster that run the Solana software (Agave).
- Protocols like turbine, gulf stream, and gossip are used to move data efficiently across that cluster.
5. Economics & tokens
TL;DR
Any decentralised network needs an automatic way to pay the people securing it. Naturally, users pay fees when they send transactions, and those fees go to validators who provide hardware, bandwidth, and security.
- SOL is the main currency the network uses for fees, staking, and rewards.
- Lamports are tiny units of SOL that let the system price things precisely.
- Staking is how SOL holders back validators with value, earn rewards, and align incentives toward honest behaviour.
6. Clients
TL;DR
Solana is the shared backend that stores state and runs programs. Most people don’t talk to the cluster directly. Instead, they use clients, RPC endpoints, and dApp frontends that turn clicks and API calls into the low‑level requests the chain understands.
- Clients are the software that connect to Solana clusters (wallets, SDKs, backends).
- RPC endpoints are the URLs that clients hit to read data or send signed transactions.
- Apps (dApp frontends) are the UIs that let users sign, send, and see what’s happening on‑chain.
7. Advanced concepts & future tech
TL;DR
Once the basics are in place, networks need new engines, better tooling, and safer consensus paths. This section covers the under‑the‑hood machinery that makes Solana faster, safer, or easier to build on, plus ideas that are still in the lab.
- Validator clients (Agave, Firedancer, Frankendancer, Jito) are alternative engines that run the same chain but with different performance and features.
- Consensus and networking upgrades (MCP, BAM, Alpenglow, Rotor, 2Z, Slashing) are new approaches to reaching agreement and moving data faster.
- Developer tooling and execution layers (SVM, Anchor, Rust) are how builders actually write and run code on top of all that infrastructure.
Conclusion
Solana can feel like an entirely new universe of jargon, moving parts, and acronyms. That’s normal: every Solana maxie, builder, and infra nerd you know started by wrestling with these same words, trying to make sense of them.
The goal of this glossary is not for you to memorise everything, but to give you a map: what each piece does, how it fits into the machine, and a simple way to talk about it without diving into the docs.