please I need a detailed information on facebook thrift seminar

facebook thrift seminar topic download in pdf

Abstract
Thrift is a software library and a set of code generation tool, which was developed in the Facebook offices in Palo Alto, California, to accelerate the development and deployment of scalable and efficient backend services.

The main objective of lean is to ensure effective and reliable communication between programming languages ​​by abstracting portions of each language, as a rule, require the most common settings in the library, which is implemented in each language. This is done by allowing users to define data types and service interfaces in the Common Interface Definition Logic File (IDL File), which is supposed to be independent of the language file and generates all the necessary code to create the remote procedure call for clients and servers, this report explains choice of design and implementation details of the level, as well as trying to demonstrate a sample service Frugality.

The whole concept of frugality stems from the fact that a new direction was required to address resource issues requires many on-site Facebook applications that can not be solved by staying within the lamp. LAMP is an acronym for Linux, MySQL, Apache and PHP. When Facebook is now painstakingly developed, it was made from scratch, using the lamp frame. By 2006, Facebook has been widely recognized throughout the world as a social networking site and, therefore, its network traffic has also increased which leads to the need to expand its network structure for many of its local applications, such as search, ad selection and delivery and registration events.

Scaling up these operations to match the resource requirements is not possible under the lamp. In their implementation the creation of many of these services, such as search, event logging various programming languages ​​has been selected to optimize for the right combination of performance, ease of development and speed of access to existing libraries, etc. In addition most of the Facebook culture has always preferred to choose the best tools and the implementation of the standards in any programming language, and one reluctantly accepting their limitations inherent. Most programming languages ​​ever suffered from subpar performance or restrict the freedom of the data type. Taking into account all these technical problems and solutions, Facebook engineers were presented with a daunting task of building scalable, transparent and high bridge over the performance of different programming languages.

Features Lean Design
The basic idea of ​​lean that it consists of a stack of neutral language, which is implemented in various programming languages ​​and the associated code generating engine which converts the simple interface definition language and data in the client and remote call procedures library server. Frugality is designed to be as simple as possible for developers who can identify all the necessary data structures and interfaces for integrated services in one short file. This file is called thrift Logic Interface Definition file or thrift IDL file. Developers have identified some important features in the assessment of the technical problems of cross-language interaction in a networked environment.

Types:
Common type system should exist in all languages ​​without requiring the need for developers to write your own serialization code. Serialization is the process of converting an object of one type to another. For example, if a programmer to write an application that implements a strongly typed STL map for Python dictionary. No programmer should be forced to write code below the application layer. The dictionary is a data type in Python, which allows the sequencing of a collection of elements or elements with the arrow keys. It is very similar to the "associative arrays".

Transport:

Each language must have a common interface for bidirectional raw data. Consider a scenario where there are 2 servers, where one is deployed in Java, and the other is deployed in Python. Thus, a typical service is written in Java, you should be able to send the raw data from the service in a common interface that is clear to another server that runs on Python, and vice versa. The transport layer should be able to transport the raw data file on the two ends. The specifics of how this is implemented, shall not have a value for the service developer. The same application code must be able to work with TCP Stream sockets, raw data files in memory or on disk.

Protocol:

For transporting the raw data, then they must be encoded in a specific format such as binary, XML, etc. Therefore, the transport layer uses some specific protocol for encoding and decoding data. Again the application developer will not bother about it. It is only if the data concerned can be read or written in a deterministic manner.

Versioning:

For services to be sustainable, they must evolve from their current version. They should include new features and to make this type of data involved in the service should provide a mechanism to add or remove or modify the fields of the object list of function arguments without any interruption in service. This is called Versioning.

Processors:

Processors are the ones that handle data streams and perform remote procedure calls.

Thrift allows programmers to develop full use of native data type of frugality, not using any shell objects or special dynamic types. It also does not require the developer to write the serialization code for transport. Developer gets freedom logically annotate their data structures in Thrift Interface Definition Logic File (File IDL), with a minimum of additional information needed to tell the code generator is safe to transport objects between languages.

: structures

Savings structure defines shared object that will be used in all languages. A structure substantially similar to a class in an object-oriented programming languages. Calculating structure has a strongly typed field with unique identifiers fields. The basic syntax for Thrift structure is very similar to the structure used in the structure of C. Fields savings can be annotated with a unique identifier field, unique for the frame structure, and additional default values. The concept of field identifiers may also be omitted, and this concept was introduced identifers field strictly for version control purposes.

Here's how to thrift Struct looks

Example structure

{
1: i32 number = 10,
2: i64 bignumber,
3: double decimal places,
4: string name = "NB"
};

As you can see the field in savings structure marked a unique identifier fields.

Facebook Thrift Services
Frugality has been used in a large number of applications on Facebook, including search, registration, mobile, ads and platform for developers. Two specific forms of application are discussed below.

Search

Frugality is used as the base protocol and transport layer for the Facebook search engine. multi-language code generation is well suited for searching, since it allows for the development of applications in an efficient server-side language (C ++) with Facebook and allows PHP-based web application calls lookup service using Thrift PHP libraries. There is also a wide variety of search statistics, deployment, and testing of functionality built on top of the generated Python code. Furthermore, Parsimony log file format is used as the journal repeats for update in real-time mode index search. Frugality helped found the team to use each language to its strengths and to develop code at a rapid pace.

logging

Functionality TFileTransport Frugality is used for structured registration. Each definition service function along with its parameters can be viewed as structured log record is identified by a function name. This log can be used for various purposes, including online and offline processing, aggregation statistics, as redo log.

Frugality has allowed Facebook to build scalable backend services efficiently, enabling engineers to divide and conquer. Application developers can focus on application code, without worrying about the sockets layer. We avoid duplicated work, writing buffering and I / O logic in one place, rather than pouring it into each application. Frugality has been used in a wide range of applications on Facebook, including search, registration, mobile, ads and platform for developers. We found that the marginal cost of productivity incurred an additional level of abstraction software is far overshadowed by improving the efficiency, reliability, and system developers. Finally Frugality was added to the Apache Software Foundation as the Apache project savings, making it an open source framework for the implementation of cross-language services.