One-stop full coverage data I/O platform - the perfect combination of Alluxio and Aunalytics

"The combination of Aunalytics' cloud-native data analytics platform and Alluxio's open source data orchestration software enables customers to have unified access across all data sources and drive AI analytics to generate better answers and gain a competitive advantage."

- Tom Panozzo, CTO, Analytic Cloud

Use Case: Unified Data Access + Local Computing Acceleration
Technology Stack : Spark + Drill + Hadoop + Custom AI & Analytics Alluxio /HDFS + NFS + Object Storage
Challenge:   Operational complexity of accessing multiple data stores + Compatibility issues of different computing services
Advantages:  Centralized access to all data + improved performance through in-memory caching + hardware flexibility with storage tiering

  • The Challenge: Finding the Digital Transformation Divide for a Single Cloud Solution

As early as 2012, as a start-up company, Aunalytics was faced with the problem of using a very complex analysis environment. As a data platform company, it aimed to provide "deep analysis as a service" to solve the most important problems for enterprises and medium-sized companies. IT and business issues. Aunalytics DaybreakTM Industry Intelligence Data Mart combines the power of its data platform to provide industry-specific data models with built-in query and artificial intelligence, ensuring timely and accurate data and answers to important business questions. Aanalytics leverages internally developed analytics software on top of open source computing technology to provide clients with rich data marts in a question-and-answer format. Any business person, regardless of technical ability, can use Daybreak products and leverage natural language processing to deconstruct real-world questions and turn them into data queries. In order to achieve the aforementioned grand vision, the Aunalytics team put a lot of effort into trying to eliminate the lack of knowledge and understaffing experienced by middle market customers in solving the so-called digital transformation gap, and the team succeeded in doing so, but as these technologies Early adopters, the team also incurs operational overhead. In 2020, Aunalytics implemented a next-generation computing platform designed to completely separate storage, compute, and delivery through adaptive analytics software (Aunsight™) for ease of development and flexible scaling. The software is capable of submitting workloads to computing services that ingest data from vast storage resources.

As a result, the Aunalytics team faced two options:

1. Centralized management on mainstream storage environments (such as NFS, iSCSI, etc.), but these environments are riddled with numerous performance, consistency and concurrency issues;

2. Employ a multi/any storage environment with a single point of access to these systems.

As an alternative to adopting Alluxio's new stack, Aunalytics evaluated a cloud computing platform that bundles storage and can be deployed in a private cloud, and finally adopted Alluxio's next-generation data orchestration system to centralize the data I/O footprint on one technology, thereby Simplifies integration with future updates to the compute engine.

  • Aanalytics uses Alluxio as a "one-stop shop" for data I/O

Alluxio provides Aunalytics' computing environment with a single point of access to all stored data, regardless of capacity, speed, or storage I/O performance. This allows the team's developers to focus on analytics integration without regard to storage environment constraints or compatibility. Alluxio is the main data access and writing mechanism for the batch computing environment, dynamic query environment, and the main product Daybreak. At the same time, Alluxio allows the Aunalytics team to leverage existing Hadoop storage, new private cloud storage, and future bulk object storage for cost-effective flexible and scalable storage. Aunalytics avoids the complexity of supporting legacy environments as the platform evolves.

Alluxio's "one-stop" service for data I/O has enabled the Aunalytics team to migrate from large monolithic analytics environments such as Hadoop while still using the hadoop storage system during the migration. Alluxio also enables teams to perform data "piloting" between analytics systems, where data can be written to Alluxio by one system and later read by a completely separate, non-compatible system.

Sharing data across systems has reduced post-migration data movement/copy needs by 90% and reduced computation and delivery times by 30% .

  • Integrate Alluxio

Alluxio was deployed as part of the new Aunalytics data environment. Alluxio enabled the Aunalytics team to do a lot of new work, but it wasn't implemented as a complementary system. The addition of Alluxio fully meets the team's requirements for flexibility and builds the next-generation Aunalytics platform on this basis, unlike other options that force the team to make trade-offs when adopting a new solution.

  • further cooperation

Alluxio fundamentally separates storage and computing, improves the speed and agility of big data and AI workloads, and enables users to migrate to newer storage solutions such as object storage, eliminating data duplication and reducing costs. Teams continue to use open source computing projects (powered by Aunsight™ and Alluxio) next to Daybreak products in the AI ​​workspace to deliver collective products.

Aanalytics provides its customers with extensive technical support for integrating Alluxio and other data management technologies, providing business users with out-of-the-box analysis solutions with built-in data management functions.

For more interesting and interesting [event information] [technical articles] [big coffee views], please pay attention to [Alluxio Think Tank] :

{{o.name}}
{{m.name}}

Guess you like

Origin my.oschina.net/u/5904778/blog/5586352