Acne Buterin: the Central element of Casper successfully tested in a test network

Meeting the developers of Ethereum are opened, however, the communication between the developers is not addressed to the General public, and so much in need of clarification.

At the last weekly meeting of the Foundation developers of Ethereum, Vitalik Buterin said:

The key concept of «Casper» Casper Protocol, which is that two conflicting block can not be finalized, works successfully on a test network, and the network finds consensus (which chain is true). This part of the test can be considered successful.

Hybrid Protocol PoS Casper FFG for several weeks running on a test network Ethereum, however, observers noted significant problems in his work. The relationship between nodes is constantly interrupted, and the network is overloaded with excessive thread connections. As a result, the network worked continuously, developers must manage it in manual mode.

Answering questions from Reddit users, acne Buterin so explained the situation:

The Casper Protocol is just a specification, it’s not code, it is not «built» on anything. Technical community Ethereum traditionally separates the specification and the implementation (implementation). Currently there is only one implementation of the specification Casper on platforms pyethereum (Casper consensus) and pyethapp (communication network), based on python, but now completed work on the implementation of Java that will be used in the new client Harmony. By that time, as Casper will be deployed on the main blockchain should be ready for implementation at the Geth and Parity.

In the code network pyethapp discovered critical errors, while implementation logic Casper tested successfully. Likely to deploy Casper in the core network in the second quarter will not be as Geth and Parity need time to develop their own implementations. And they, in turn, should begin only after the final round of the critical elements of the Protocol (namely, partial fines and transactions related to the duration of the sessions of the validators).

This means that require additional testing before launch. Still, the developers hope that FFG Casper will be ready by the end of summer.

As for sharding, which will be the basis of future technology of Ethereum, in this direction, according to Buterin, «began the big job.»

Currently, every node in the blockchain have to do the same amount of computation, a this means that the total computing power of the network is limited by the capacity of its weakest node. Sharding introduces parallel computing: instead of having to reproduce all of the transaction nodes into groups (shards), for example, 100 nodes each, and each shard will sign 1000 transactions.

Then, using the algorithm of consensus, all these nodes are linked together so that if any node is pursuing the wrong transaction, the whole network receives the information and rejects it. So the developers are going to allow scalability of the system, in this case, as was said at the meeting, work on the first part of the first phase of the project, namely, specification, are at the proper level. This phrase has caused confusion in some media, which hastened to declare about completion of work on the first of the four phases of the project.

Soon follow
the clarification from the developer Fund nick Johnson (Nick Johnson):

[In the work plan for sharding] there will be 4 phases in the long term and 4 stages in the first phase in the near future. The first phase is almost ready and testing plus some work on the integration will be completed within a month. (In parallel we will begin work on the second stage).

Meanwhile, the current network of Ethereum’s still working at the bandwidth limit, although transaction fees fell to normal – about 10 cents. Until now, the framework developers paid more attention to the sharding and Casper, and works to optimize the client software was a lower priority. Now the situation has changed and reduced commissions, albeit temporary, was the result of such work. Peter Ilahi (Péter Szilágyi), the developer Geth – one of the clients, illustrated work optimized client Geth v1.8.0 (not released yet), compared to the operated v1.7.3.

While the Fund Ethereum is looking for ways of overcoming «bottlenecks», even the optimization of existing software is able to show impressive results.

Добавить комментарий