Skip to content

VaagaTech · Graph Engine

Distributed Graph Intelligence Optimized for Cloud-Native Density

Designed for modern containerized environments, VaagaGraph enables high-volume graph processing within modest resource allocations. By pairing consistent hash sharding with tiered caching and quorum-verified persistence, it delivers responsive traversals and flexible match threshold evaluations with dependable data durability.

Enterprise Core Capabilities

Thoughtfully architected for cloud-native reliability

Designed to maintain predictable memory boundaries, operate without heavy runtime overhead, and evaluate multi-hop relationship affinity with precision.

Dynamic Match Threshold Querying

Unlike conventional strict binary filters, VaagaGraph supports configurable percentage threshold evaluation. Query by custom affinity thresholds (such as 25%, 50%, or 75%), weighted multi-attribute criteria, and connection requirements to receive ranked candidate results with sub-10ms responsiveness.

📦

Predictable Container Memory Footprint

Tailored for Kubernetes memory boundaries. An active 75% resource governor reserves headroom for runtime garbage collection, while tiered LRU caching and sparse indices ensure consistent operational stability under intensive query loads.

🛡️

Synchronous Quorum Durability

Synchronous write-ahead log replication across distributed replicas, paired with optimistic concurrency controls on durable storage, ensures continuous availability and rapid failover recovery.

🧩

Unified Monorepo Architecture

Available as an embeddable Node.js engine (@vaagatech/vaaga-graph), a distributed EKS database daemon, a client SDK (@vaagatech/vaaga-graph-sdk), and an interactive Studio management UI.

💰

Cost-Efficient Resource Scaling

Avoid the sustained overhead of dedicated idle instances. VaagaGraph compute scales gracefully with actual demand, while persistent graph state rests securely on durable object storage.

Sub-10ms Traversal Speed

Dual adjacency index maps (in/out edges), Bloom filters, and inverted label indices deliver instant traversal latency without native C++ compilation bindings or JVM warmup delays.

Developer Experience

Dynamic Match Threshold Query in Action

Find candidate nodes matching any custom percentage threshold (e.g. 25%, 50%, 75%, 90%) with multi-attribute weighted scoring:

import { VaagaGraphClient } from '@vaagatech/vaaga-graph-sdk';

// 1. Connect to distributed cluster (1GB EKS pods)
const client = new VaagaGraphClient({
  endpoints: ['http://vaaga-graph.default.svc.cluster.local:4000']
});
await client.connect();

// 2. Execute Dynamic Match Threshold Query across shards (e.g. 50% or custom)
const { results, latencyMs } = await client.query().match({
  label: 'Engineer',
  threshold: 50, // Pass ANY custom threshold: 25, 50, 75, 90...
  props: { role: 'Architect', city: 'Chennai' },
  ranges: { experienceYears: { min: 5 } },
  connections: [
    { type: 'WORKS_AT', targetLabel: 'VaagaTech', weight: 2 }
  ]
});

console.log(`Matched ${results.length} nodes in ${latencyMs}ms`);
results.forEach(r => console.log(`${r.node.props.name}: ${r.matchPercentage}% match`));