Hadoop： Hadoop 是一个用Java编写的开源软件框架，用于存储数据和处理从千兆字节到 PB 级的大型数据集。 Hadoop 是一种分布式文件系统，可以跨计算机存储和处理海量数据集群。开源的 Hadoop 与所有平台兼容，因为它是基于 Java 的。 Hadoop 有两个核心层，即处理/计算层(MapReduce)和存储层(Hadoop 分布式文件系统)。 Hadoop 跨计算机集群运行代码，并对跨商用服务器集群的庞大数据集执行离线批处理。但是，Hadoop 不是 SQL 的替代品，而是它们的使用取决于个人需求。在性能方面，Hadoop 优于 SQL，因为它具有更高的速度和以相同效率处理结构化、半结构化和非结构化数据的能力。
SQL 性能：结构化查询语言 (SQL) 是一种用于在数据库中操作、检索和存储数据的标准语言。关系数据库使用 SQL 作为标准来维护和操作数据。 SQL 命令，例如“选择”、“插入”、“更新”、“删除”、“创建”和“删除”，可用于存储、更新或检索数据库中的数据。一些使用 SQL 的常见关系数据库管理系统有 Oracle、Microsoft SQL Server、Sybase、Access、Ingres 等。 然而，随着数据量(或大数据)的增加，使用 SQL 存储如此大量的数据变得困难。关系数据库。对于结构化模式效果很好，但对于大数据，它没有固定模式，而是半结构化数据。 RDBMS 大数据的 3 V：容量、多样性和速度是导致 NoSQL 数据库出现的主要原因。从名称上看，很明显 SQL 不能再用于 NoSQL 数据库的数据操作。在这种情况下，Hadoop 比 SQL 更具优势。
下表列出了 Hadoop 和 SQL 性能之间的差异：
| Feature | Hadoop | SQL Performance |
|---|---|---|
| Structure | No fixed schema | Fixed Schema |
| Data Format | Structured, semi-structured or unstructured data | Structured data |
| Data Volume | Hadoop works exceptionally well on both low and high volume of data | SQL works better on low volume of data |
| Data processing | Hadoop supports large-scale offline batch processing known as OLAP | SQL supports Real-time data processing known as OLTP |
| Speed | Faster | Slower |
| Throughput | Higher throughput | Lower throughput |
| Latency | Hadoop cannot fetch a particular record from the data set very quickly hence it has low latency | SQL can fetch a particular record from the data set very quickly hence it has high latency |
| Scalability | Horizontal scalability which means more machines can be added in the network for parallel processing | Vertical scalability which means more hardware or CPU is added to existing machine |
| Data Storage | Data can be stored in the form of tables, key-value pairs etc | Data can be stored in the form of tables only. |
| Integrity | Low integrity | High integrity |
| Data variety | Hadoop deals with Big data and supports variety of data | SQL does not support variety of data |
| Updates | Hadoop is designed with the concept of write once read many. Hence data updates are practically not possible | SQL is write once, read and update many. Hence data updates are very easily done |
| ACID Properties | It does not fully comply with ACID properties | It fully complies with ACID properties |
| License | Hadoop is free open source software | SQL is licensed |
| Example | MongoDB, HBase etc | Oracle, Microsoft SQL Server etc |