The Hydrolix platform
Architected for massive scale without compromises on data quality, query performance, retention, or cost. Learn about the engine that gives you both real-time analytics and long-term insights on petabytes of data.
Real-time analytics. Full-fidelity data.
No compromises.
Stream data in real time, even during peak events generating tens of millions of events per second.
Transform and optimize data before storage. Includes standardizing, enrichment, and masking.
Store and retain data long-term in object storage. Cost-effective, high-availability data with massive compression and no data tiering.
Search data in real time regardless of age, with all data hot for queries. Use MCP to enable natural-language queries with AI assistants.
Designed for the biggest challenges in data
High-volume streaming ingest
Ingest tens of millions of log lines per second in real time, with enterprises typically ingesting at least 1 billion log lines per day.
Long-term retention
Retain data 15 months by default. High-density compression and object storage allow you to keep petabyte-scale data for a fraction of the cost of other solutions.
Hot data
Query in real time on all your data regardless of age. There are no data tiers to manage or performance limitations for historical data.
Full-fidelity data
Retain full-fidelity datasets, not just aggregations or sampled data.
Cost-optimized
Reduce costs compared to other solutions. Cost-optimized so you can keep all your data without compromises.
Full control
Keep data in your own storage for complete control or bring your own cloud (BYOC) and run Hydrolix fully in your infrastructure.
One platform. Every cloud. All your tools.
Fully managed or BYOC
- Choose fully managed or bring your own cloud (BYOC). Store your data in your object storage with no vendor lock-in or data egress.
- Run Hydrolix infrastructure (with BYOC) in your own virtual private cloud (VPC) or in multiple clouds for greater control.
- Compatible with all major clouds.
Integrations and compliance, available on day one
- Use the visualization tools you prefer, including Grafana, Kibana, Superset, and Looker.
- Integrate with the Apache Spark ecosystem, including Databricks, AWS EMR, and Microsoft Fabric.
- SOC 2 and GDPR compliant, with granular role-based access control, row and column control, and strict separation between projects for increased data security.
- Ingest telemetry from any source compatible with OpenTelemetry (OTel) using native OTLP support.
Platform Benefits
Architected for performance and scale, not skyrocketing costs
Traditionally, enterprises have relied on costly, vertically-scaled hardware to deliver real-time analytics. For big data, this approach is challenging to scale and too expensive. Hydrolix combines the power of modern cloud computing with an engineering approach that maximizes the performance of distributed object storage.
Decoupled architecture
All components are stateless and decoupled, allowing each subsystem to scale independently and without resource contention. For example, ingest scales to handle peak events. Meanwhile, you can scale up query during urgent investigations.
Massive parallelism
All components use massive parallelism to maximize the benefits of cloud computing. For example, a Hydrolix cluster can scale to hundreds of intake heads, all writing partitions in parallel for major events.
Columnar storage
With columnar storage, you can query individual columns, leading to more efficient queries than row-based storage. Columnar storage also makes it possible to compress columns individually for greater compaction.
Advanced compression
Advanced algorithms optimize compression for each column individually based on data type and other factors. Compression rates are typically 20x-50x, leading to faster read and write times and lower storage costs.
Merge service
Partitions are small at ingest time (resulting in faster time to insights). Over time, an automated merge service runs in the background and optimizes partitions for improved compaction and query performance.
Block-level indexing
All components autoscale individually to meet demand, ensuring your infrastructure remains efficient. You can also manually scale to ensure optimal performance during major events or even scale down to zero to reduce compute and costs during off-peak times.
Decoupled object storage
Data is written to object storage. In addition to being cost-effective, it is highly scalable for big data and long-term retention.
Streaming ETL
Real-time streaming and data transformation (streaming ETL) optimizes data for storage. This process includes standardizing, compressing, and partitioning data for optimized query performance.
Summary tables
Summary tables store real-time aggregations and metrics separate from raw data tables. Summary tables are updated as data is ingested, ensuring they remain highly accurate.
Time-based partitioning
All data is partitioned by time, using partition pruning to make time-based queries efficient.
Scaling
Transforms serve as schemas, defining how incoming data should map to system columns and how it should be stored.
Immutable, append-only data
Data is immutable and append-only, ensuring data integrity and leading to faster reads and writes.
Customer Story
Real time analytics at super scale
Traditionally, enterprises have relied on costly, vertically-scaled hardware to deliver real-time analytics. For big data, this approach is challenging to scale and too expensive. Hydrolix combines the power of modern cloud computing with an engineering approach that maximizes the performance of distributed object storage.
RESULTS
17.4 GB
SECOND PEAK DATA INGEST RATE
5 to 10 seconds
time to glass
55,000
QUERIES OVER COURSE OF EVENT
16x
RAW DATA COMPRESSION RATES
98.4%
row reduction
0.481 second
QUERY RESPONSE TIME
Talk to an expert about your use case
When you need faster insights on big data, from edge to enterprise. Full-fidelity data, no compromises.