Aptos is a revolutionary solution to the scaling problem!
Blockchain technology has many advantages: it increases the level of security, makes it easy to conduct transactions, and keeps a record of all transactions. However, as its use becomes more widespread, a any problems arise. One such issue is scalability.
In blockchain technology, the term "scaling" describes the increase in system throughput, measured by the counts of transactions performed per second.
The throughput of the Bitcoin blockchain system is only 3 to 7 TPS; Ethereum ~ 15 TPS.
With the constant increase in the use of cryptocurrencies, it became necessary to create "layers" to improve network security, record keeping and much more.
Layer 1 in a decentralized ecosystem is the blockchain itself. Layer 2 is a third-party add-on to layer 1 that increases the number of nodes and therefore the throughput of the system.
Level 2 takes over part of the transaction load of the first level of the blockchain and transfers it to another system architecture, effectively offloading the system with an influx of transactions.
Since most of the processing load falls on this adjacent supporting architecture, the load on the rest of the network decreases, and therefore the layer 1 blockchain becomes not only less overloaded, but also more scalable.
Task
Aptos set themselves the task get design a parallel execution engine that extracts the maximum inherent speedup that is possible, given the actual access conflicts in the workload. For the best programming and user experience, the algorithm should be transparent to the users. An alternative approach taken by some blockchain parallel execution engines is to force the user to declare the dependencies upfront, which severely limits what a transaction can do and may require a transaction to be broken up or retried. Instead, to avoid incurring such costs and programming annoyances, a system design goal for our parallel execution engine is to manage all conflicts internally and automatically adapt to the workload.
Decision
Aptos has found a revolutionary solution to the problem!
In developing Block-STM, management paid special attention to the observation that traffic use case:
- No need to commit transactions individually;
- VM provides safety for optimistic memory access;
- Pre-defined order reduces synchronization.
Block-STM combines known techniques with novel ideas:
- Optimistic concurrency control;
- Multi-version data structure;
- Validation;
- Collaborative schedule;
- Dynamic dependency estimation.
Evaluation
Aptos Implemented Block-STM in safe Rust in the Aptos open source codebase, relying on Rayon, Dashmap, and ArcSwap crates for concurrency. Aptos evaluated the system with non-trivial peer-to-peer Move transactions (8 reads and 5 writes). In the graph below we compare Block-STM against a sequential execution of the block. Every block contains 10k transactions and the number of accounts determines the level of conflicts and contention. Under low contention Block STM archives 16x speedup over sequential execution with 32 threads, while under high contention Block-STM archives over 8x speedup. Importantly, Block-STM incurs a small overhead when the workload is inherently sequential. Overall, Block-STM is able to dynamically and transparently (without any hints from the user) extract the inherent parallelism from a workload.
Aptos has thus created the most secure and scalable L 1 blockchain.
You can read more here.
Website | Medium | Twitter | Discord | Telegram
