Scale Look-to-Book Ratio for Online Travel Agent System

Situation

A large international tour operator (the “operator”) provides 24-hour services to its travel agents and Internet shoppers on HPE NonStop Servers. Utilizing the Internet, the tour operator’s business grew and look-to-book ratios increased several-hundred fold (from 30:1 to 700:1).

Problem

1. The search query volume carried such a heavy load (up to fifty times the original query load) that the system ran out of capacity.
2. The operator recognized that booking activity evolved into approximately 99% complex queries and 1% simple writes.
3. They needed to scale-out the query system to a multi-node, distributed solution based on commodity hardware.

Solution

• Use HPE Shadowbase Data Integration for integrating heterogeneous systems in an asymmetric capacity expansion.
• Scale query load capacity in a master-subordinate relationship, replicating from the HPE NonStop booking system to both NonStop and Windows query “looking” systems.

Additional Asymmetric Capacity Expansion (ACE) Architecture Notes

• ACE is similar to a “master-subordinate” architecture.
• The master-booking node is hosted on an HPE NonStop Server, which is connected and synchronized with any number of scalable NonStop and Windows query nodes via HPE Shadowbase data replication.
  • The key structure on the master-booking node remains optimized for online transactional processing (OLTP) access (“skinny” keys), and the key structure on the query node is optimized for query (“fat” keys).
  • Data from the master-booking node is also replicated to Windows servers running SQL, which supports a large Operational Data Store (ODS) application.
• This read-only function is offloaded from the master-booking node to the commodity Windows hardware, since travel agents and Internet travel companies use the query nodes to look-up vacancy, seat availability, etc.
• The outsource provider’s application is configured (no application modification was required) to perform all queries against the query nodes, yet still apply the booking transactions to the master-booking node.

Outcomes

• Scales queries to handle varying (and very large) query volumes
  • The entire database resides on each of the read-only nodes, essentially allowing the provider to scale the application across multiple systems to handle unpredictable query volumes
  • Additional commodity hardware can be added as query nodes to scale with increasing query load
  • Reduces Total Cost of Ownership (TCO) by leveraging commodity systems for Internet queries
• Increases availability with characteristics similar to an active/active architecture
• Introduces Disaster Recovery capability for the NonStop master node
  • If the NonStop master-booking node fails, the “look node” can act as a backup

HPE Shadowbase Products of Interest

• HPE NonStop Shadowbase Basic Data Integration Software (BE443AC/QSA51V6)
• HPE Shadowbase Basic Application Software 1-8 core or 9+ core (WSA51V6T1/T2)

Contact us or your HPE Shadowbase representative, and learn how Shadowbase software will benefit you.

Further Reading

Related Case Study: Large International Tour Operator Uses HPE NonStop to Optimize Look-to-Book Processing

Related White Paper: HPE Shadowbase Streams for Data Integration

Related Solution Brief: HPE Shadowbase Data Replication Solutions