Internet Computer Overview

The Internet Computer (IC) is a decentralized cloud platform. Actually, it is better thought of as a progressively decentralizing cloud platform. Its full vision is yet to be fulfilled.

互联网计算机（IC）是一个去中心化的云平台。 实际上，最好将其视为一个逐步去中心化的云平台。 它的全部愿景尚未实现。

It aims to be owned and operated by many independent entities in many geographies and legal jurisdictions throughout the world. This is in opposition to most traditional cloud platforms today, which are generally owned and operated by one overarching legal entity.

它的目标是由世界各地许多地区和法律管辖区的许多独立实体拥有和经营。 这与当今大多数传统云平台相反，后者通常由一个总体法律实体拥有和运营。

The IC is composed of computer hardware nodes running the IC protocol software. Each running IC protocol software process is known as a replica.

IC由运行IC协议软件的计算机硬件节点组成。 每个正在运行的 IC 协议软件进程称为副本。

Nodes are assigned into groups known as subnets. Each subnet attempts to maximize its decentralization of nodes according to factors such as data center location and node operator independence.

节点被分配到称为子网的组中。 每个子网都试图根据数据中心位置和节点运营商独立性等因素最大化其节点的去中心化。

The subnets vary in size. Generally speaking the larger the size of the subnet the more secure it will be. Subnets currently range in size from 13 to 40 nodes, with most subnets having 13 nodes.

子网的大小各不相同。 一般来说，子网的规模越大，它就越安全。 当前子网的大小范围为 13 到 40 个节点，大多数子网有 13 个节点。

IC applications, known as canisters, are deployed to specific subnets. They are then accessible through Internet Protocol requests such as HTTP. Each subnet replicates all canisters across all of its replicas. A consensus protocol is run by the replicas to ensure Byzantine Fault Tolerance.

IC 应用程序（称为容器）部署到特定的子网。 然后可以通过 Internet 协议请求（例如 HTTP）访问它们。 每个子网都会在其所有副本中复制所有容器。 副本运行共识协议以确保拜占庭容错。

View the IC Dashboard to explore all data centers, subnets, node operators, and many other aspects of the IC.

查看 IC 仪表板以探索 IC 的所有数据中心、子网、节点运营商和许多其他方面。

Canisters Overview

Canisters are Internet Computer (IC) applications. They are the encapsulation of your code and state, and are essentially Wasm modules.

容器是互联网计算机 (IC) 应用程序。 它们是代码和状态的封装，本质上是 Wasm 模块。

State can be stored on the 4 GiB heap or in a larger 64 GiB location called stable memory. You can store state on the heap using your language's native global variables. You can store state in stable memory using low-level APIs or special stable data structures that behave similarly to native language data structures.

状态可以存储在 4 GiB 堆上或更大的 64 GiB 位置（称为稳定内存stable memory）。 您可以使用语言的本机全局变量将状态存储在堆上。 您可以使用低级 API 或行为与本机语言数据结构类似的特殊稳定数据结构将状态存储在稳定内存中。

State changes must go through a process called consensus. The consensus process ensures that state changes are Byzantine Fault Tolerant. This process takes a few seconds to complete.

状态改变必须经过一个称为共识的过程。 共识过程确保状态更改具有拜占庭容错能力。 此过程需要几秒钟才能完成。

Operations on canister state are exposed to users through canister methods. These methods can be invoked through HTTP requests. Query methods allow state to be read and are low-latency. Update methods allow state to be changed and are higher-latency. Update methods take a few seconds to complete because of the consensus process.

对容器状态的操作通过容器方法向用户公开。 这些方法可以通过HTTP请求来调用。 查询方法允许读取状态并且延迟低。 更新方法允许更改状态并且延迟较高。 由于共识过程，更新方法需要几秒钟才能完成。