Solution Architecture

The overall architecture of Kunpeng BoostKit for Big Data consists of the hardware platform, OS, middleware, and big data platform. The supported big data platforms include Huawei-developed FusionInsight, open source Apache, Transwarp, and China Mobile Suzhou Research Institute big data platforms.

Figure 1 shows the overall architecture of Kunpeng BoostKit for Big Data.

Figure 1 Overall architecture of Kunpeng BoostKit for Big Data

**Table 1** Components
Component	Description
Hardware	Kunpeng servers powered by Huawei Kunpeng processors are used. SSD acceleration is supported for high-speed cache scenarios.
OS	Mainstream commercial and open source OSs are supported.
Acceleration	Erasure Code encoding and decoding are accelerated by using Kunpeng instructions. The processor built-in encryption and decryption accelerator provides a KAL machine learning acceleration library for algorithm and instruction optimization.
Compiler	Common big data compiler middleware such as JDK, GCC, and LLVM support the Kunpeng industry ecosystem.
Big data component	Open source big data components, including Hadoop, HBase, Spark, Hive, Flink, and Elasticsearch, are supported.
Big data platform	Huawei-developed FusionInsight, open source Apache, Transwarp, and China Mobile Suzhou Research Institute big data platforms are supported.