13.9 Answers to questions from Section 13.5 回答第 13.5 节中的问题

Does a developer need to know each of the following pieces of information in order to use the IndexLookup class?

开发人员是否需要了解以下每条信息才能使用 IndexLookup 类？

1. The format of messages that the IndexLookup class sends to the servers holding indexes and objects. No: this is an implementation detail that should be hidden within the class.
2. The comparison function used to determine whether a particular object falls in the desired range (is comparison done using integers, floating-point numbers, or strings?). Yes: users of the class need to know this information.
3. The data structure used to store indexes on servers. No: this information should be encapsulated on the servers; not even the implementation of IndexLookup should need to know this.
4. Whether or not IndexLookup issues multiple requests to different servers concurrently. Possibly: if IndexLookup uses special techniques to improve performance, then the documentation should provide some high-level information about this, since users may care about performance.
5. The mechanism for handling server crashes. No: RAMCloud recovers automatically from server crashes, so crashes are not visible to application-level software; thus, there is no need to mention crashes in the interface documentation for IndexLookup. If crashes were reflected up to applications, then the interface documentation would need to describe how they manifest themselves (but not the details of how crash recovery works).

1. IndexLookup 类发送给包含索引和对象的服务器的消息格式。否：这是应隐藏在类中的实现细节。
2. 用于确定特定对象是否在所需范围内的比较功能（使用整数，浮点数或字符串进行比较吗？）。是：该课程的用户需要了解此信息。
3. 用于在服务器上存储索引的数据结构。否：此信息应封装在服务器上；甚至 IndexLookup 的实现都不需要知道这一点。
4. IndexLookup 是否同时向多个服务器发出多个请求。可能：如果 IndexLookup 使用特殊技术来提高性能，则文档应提供有关此问题的一些高级信息，因为用户可能会在意性能。
5. 处理服务器崩溃的机制。否：RAMCloud 可从服务器崩溃中自动恢复，因此崩溃对于应用程序级软件不可见；因此，在 IndexLookup 的接口文档中无需提及崩溃。如果崩溃反映到应用程序中，则接口文档将需要描述它们如何表现出来（而不是崩溃恢复如何工作的详细信息）。