Blockspace: The Product of Polkadot

Blockspace

In this lesson, we will cover:

Notes:

Deep Dive Article is https://www.rob.tech/polkadot-blockspace-over-blockchains/

Blockspace: Definition

Blockspace is the capacity of a blockchain to finalize and commit operations.

Measuring Blockspace

Blockspace can be measured in a few different ways:

Blockspace Markets

Fee markets in blockchains are examples of blockspace markets.

The blockchain sells blockspace on-demand, and users pay fees in order to utilize the blockspace.

Blockspace Markets

Parachain slot auctions are another example of a blockspace market.

Instead of selling blockspace on-demand, blockspace is sold in bulk up-front with an auction mechanism.

Blockspace: Supply and Demand

Blockspace in Ethereum: GasToken

GasToken (https://GasToken.io) was an early Blockspace Futures Market on Ethereum.

Ethereum provides a gas refund for storage slots being cleared in smart contract execution.

By "pre-buying" storage when gas is cheap and getting the refund when gas is expensive, users can perform blockspace arbitrage in Ethereum!

Evaluating Blockspace

3 Properties of Blockspace:

Polkadot's Blockspace: Quality

Polkadot's Execution Sharding guarantees that all blockspace generated by Polkadot is highly secure, with economic guarantees of finality under the 33% BFT assumption.

Polkadot's Blockspace: Availability

With sharding, Polkadot has the capability to produce large amounts of blockspace. This is another lens to view the blockchain scaling problem through: creating more blockspace.

Polkadot's Blockspace: Flexibility

Polkadot provides blockspace in a highly flexible format due to key design choices:

Substrate: The Blockspace Transformer

Since Polkadot provides highly flexible blockspace, it can be transformed into a variety of different, more specialized blockspace products.

Principles of Blockchain Application Development

Notes:

By (3) I mean that the amount of demand for (2) should inform the amount to which the application does (1).

Problem: Ghost Chains

It's quite common for chains to produce mostly empty blocks.

This is a problem: chains are buying more blockspace than they need!

They are paying validators to do nothing of value, and this will lead to depreciation of their token.

Solution: Acquire Blockspace on-demand

Blockchains as needing to produce blocks every X seconds or minutes.

Blockchains should only produce blocks when they have a good reason to.

The main reason this is not done is because there are no good primitives for it.

Polkadot's Architecture: Execution Cores

Notes:

Cores, by metaphor, are like CPU cores. Code and data are scheduled onto them by the "Operating System" and then executed.

Polkadot's Architecture: Execution Cores

Long-term vs. on-demand

On-demand are analogous to "spot" instances and slot auctions are analogous to "reserved" instances in cloud computing.

Spot instances may be more expensive if overall demand is high, but help to soothe load.

Coretime

Within Polkadot, we measure the amount of blockspace that an application can use in coretime.

Just like a CPU and OS, there is a scheduler that multiplexes many
different processes onto execution cores.

Coretime is acquired through either primary or secondary marketplaces.

Elastic Scaling (planned upgrade)

What if parachains could acquire not just one execution core at a time, but multiple?

Parachains would then be able to elastically scale during periods of higher demand.

Notes:

This scaling can occur as a result of the property that in Polkadot, parablock state transitions are completely encapsulated and validation of block X+1 can occur in parallel with validation of block X.

However, the parablocks still must be authored sequentially by collators, and for that reason this can only be used to scale up to the maximum throughput of collators authoring blocks.

Reimbursement Layer: Status Quo

Reimbursement Layer: Generalizing

Notes:

User could be a collator itself, or perhaps just someone fulfilling a market need (blockspace arbitrage!)

Reimbursement Layer: Use Cases

Blockspace and Interoperability

In interoperable blockchain applications, the application is only as good as the weakest chain it relies upon.

It is important not to mix high-quality blockspace with low-quality, due to toxicity risks.

Ephemeral Chains

Is there any reason a blockchain should run forever?

Why not create blockchains that run for a limited period of time, e.g. to run some specific protocol or computation, and then conclude?

Coretime Futures Markets

With the right core-level primitives, it will be possible to transfer claims on future coretime.

Secondary markets can emerge, perhaps using NFTs, on parachains themselves, to facilitate the market for future coretime.

This will create an efficient market and price discovery for coretime via arbitrage.

Future Coretime Allocation Mechanisms

If coretime is the "product" of Polkadot, then allocation mechanisms are the "packaging".

RFC-1 proposes mechanisms for bulk coretime to be sold off, renewed, split up, resold, and transferred.

Blockspace: Conclusions

Questions