September 8, 2017 Leave a Comment
Avatar
Inside Cisco IT

AQA: A Faster Way to Index, Search, and Retrieve Large Volumes of Application Data

It’s an ugly domino effect: As application data grows exponentially—and more rapidly—by the day, the vertical stability of relational database management systems (RDBMS) saturates, and query performance degrades. These conditions undermine the performance of applications that rely on those databases, which ultimately hurts users, customers, and the business.

Cisco IT has developed a solution called Application Query Acceleration (AQA) to help Cisco’s business team make it easier for customers to index, search, and retrieve large volumes of data faster and more cost-effectively. The platform is architected to scale and is a high-performing and ready-to-use service. Designed for data-oriented applications, AQA helps retrieve constrained data in large volume datasets, with 10X faster performance compared to a database.

AQA versions, features, and capabilities

To date, Cisco IT has released two versions of AQA, both created using open-source technology. We built the base version, Version 1.0, on Apache Lucene. Version 2.0 of AQA, built on Apache Solr, has been implemented at Cisco as a cross-functional platform. It has several more advanced features, including the ability to be more scalable, recoverable, secure, distributable, and optimizable.

The AQA solution indexes database tables and views as flat files, and slices, dices, and distributes indexes across multiple nodes for performance. It also distributes search query execution to the nodes, collects datasets, and aggregates and returns the result set in less than 500 milliseconds. In addition, AQA syncs indexes with database updates.

Other features and capabilities of AQA include:

  • A standard web service interface (XML/JSON)
  • P1-type stability, security, scalability, and availability—24×7
  • Query, joins, filters, sorts, and pagination
  • Ease of administration via an administrative GUI (config/manage)
  • Supported operations: READ & WRITE, CRUD
  • Load balancing and security
  • Monitoring and alert mechanism for the service
  • Declarative indexing based on the data models
  • Disaster recovery-enabled

Version 2.0 of AQA is versatile and easily adaptable, and can use other search and NoSQL engines. It also can be integrated easily with other data analytics solutions within Cisco, such as the Cisco Log Intelligence Platform (CLIP). AQA’s technology and systems are also centralized and can be ported easily to cloud. The solution is designed to be flexible so more customized features, maintenance, and support can be added in the future.

Major business groups across Cisco IT have already integrated AQA as part of their databases, helping their customers to search and retrieve information more efficiently. So far, AQA has processed more than 220 million queries addressing the needs of these customers.