Dataswft enables sub-second interactive queries on large scale machine data (100's of Terabytes) stored across In-memory (DRAM), SSD, HDD, and Objectstores

Machine data not just comes in at great speeds adding up to huge volumes very rapidly, but is also very different from other data types such as in structured transactions or unstructured text & media. 

To harness the opportunity, and capture the value in a split-second, Dataswft delivers sub-second results for analytical queries. Machine learning and Interactive queries on 100's of Terabytes data/billions of Records comprising machine data such as readings from sensors, wearables, clickstreams, server logs, etc. both current and historical is possible with Dataswft in your IT stack. 

Dataswft is a Machine Data Lake - a centralized reservoir for all your machine data with its unique Indexing architecture that reduces dependency on a solely In-Memory processing. Dataswft allows queries to go beyond datanodes, extending to Objectstores such as S3 storage. 

Dataswft is developer friendly and readily integrates into your big data stack to boost performance several fold. Dataswft in analytics applications complements Spark, Hive and Tez to further speed up query response times.

What it means for our customer's data scientists and developers

1% Energy savings from better Building Management analysing sensor data measuring 500TB 

Creating a multi-million dollar Customer Data Platform analysing clickstream data of over 60TB in sub-seconds. 

50% reduction in manual effort by predicting likely faults at over 90% accurately using advanced machine learning algorithms such as iSAX to prevent Telecom equipment failures

Why Dataswft is well-suited for machine data?

1. Deals with Non repeating sensor readings and Sensor blank spaces  

2. Along with readings, keeps text description (like alerts) or names to find a keyword inside it. Supports keyword searches.  

3. Deals well with data in hundreds of columns and complex Boolean filters.  

4. Operations include Averaging sensor readings (1min/5min/15min averages) 

5. Read-optimized for Write once workloads

Dataswft for Better, Faster, Cheaper results

Dataswft is a key technology that complements our clients ability to serve their customers with improved operational efficiencies and enhanced customer experiences. Its cheaper as a store for all machine data, enables faster analytics, better predictions that are more accurate.

Dataswft Machine Data Lake - for better, faster and cheaper results

Few of our esteemed clients


Incident Prevention

Predicting and Preventing Telecom Tower incidents

A Telecom Tower Management company was seeking to flip its approach from reactive to proactive management, routinely analyzed network element failure incidents. They turned over their raw data from sensors and incidents spanning 1.5 years.

Dataswft powered advanced machine learning leveraged this historical data to learn and predict over half the incidents with 90% accuracy i.e. very low false positives.

•50% reduction of manual effort in monitoring

•Optimized shift planning based on predicted work loads

Besides existing rule based handling of incidents, undetected, anomalous incidents occur in significant volume requiring a manual monitoring and intervention. Current system raises tickets post incident and assigns staff to rectify failures. Tower management company seeks to be able to predict and prevent these events proactively

Real-time Targeting

Omni-channel Marketing Automation

Omni-channel marketing automation product is required to match audience to the marketers KPIs thr and place display ads using realtime clickstream analysis. 

Existing Hadoop based solution is unable to deliver query response times in seconds.

Dataswft deployed on 60TB of clickstream data demonstrated response times of 1 - 2 seconds, with 30X improvement in performance using just1/5th of existing compute infrastructure.

Dataswft is a cost effective and high performance solution providing a key differentiator to compete with large incumbents in Digital Marketing universe to achieve multi-million dollar growth.

Energy Savings

Energy Savings for Building Management

A Fortune 100 company in building energy management needs to analyse 500TB of sensor data in realtime to achieve Save Energy goal. 

Existing solution deployed on a enterprise data warehouse was unable to deliver the desired performance of under 5 seconds. Response time was in the order of several minutes.

Dataswft demonstrated consistent performance of 50 milliseconds to 5 seconds providing a 60X improvement in response times while meeting the business critical requirements.

Dataswft based Solution has created the necessary business viability the company wanted to market a new energy efficient building management solution

ACCURATE PREDICTIONS: Upper segment shows Prediction Dashboard - Live, Lower segment shows Prediction Scores Confidence spread by probability