Goals

• Improving performance of single Tephra Transaction Server
• Make Tephra Transaction server scale horizontally
• Make Tephra Transaction server Highly Available (HA) with Isolation
• Improve operational aspects of Transaction Server
• Scale the ability of system to run 1000s of WorkflowsImprove performance of Workflow scheduling to schedule 1000s of jobs / second

Areas of Focus

• Efficient Management of Transaction Invalid List Management
• Tephra Single Tephra Manager performance improvementScaling and Isolating Tephra Manager Server Performance Improvements
• Isolation architecture of Tephra
• Improving scheduling performance of Workflow system

High Level Requirements

Transaction Invalid List

Management

• System should automatically handle pruning of the transaction invalid transaction list for Major compaction
• Pruning the list during striped compactions

• list
• Reduce operational complexity for running manual steps to prune invalid transaction list
• Applied during major and striped compaction
• Metrics around the current invalid list size
• Tool to inspect the invalid list

Tephra Performance Improvement

• Single Tephra Server should be able to support up-to 10K transactions/second
• Reducing the lock granularity during conflict detection
• Supporting read-only transactions
• Improving group commit efficiency
• Support for Hierarchical conflict detection
• Support for transmitting only latest snapshot of Invalid list

Scaling

Technical Breakdown

P&S-001

: Conflict Detection Lock Granularity Improvements

P&S-002 : Read-only Transactions

P&S-003 : Improving group commit efficiencies

P&S-004 : Hierarchical conflict detection

P&S-005 : Transmitting latest snapshot of Invalid List

P&S-006 : Invalid List Pruning with major compaction

P&S-007 : Invalid List Pruning with striped compaction

P&S-008 : Tool to inspect Invalid List