Remodelling Workflow to escalate Efficiency

Sahana Systems in-house team has worked with a US-based GPS provider that is into the business of serving business instrumentation, agriculture, transportation, mobile resource management, and fleet management.


  • Our solution to their concern was simple yet productive: Big Data Analytics
  • Technology: MongoDB
  • Integrations: GPS Devices

Key Features:

  • Data compilation of 1.5TB/day from 0.5-0.75 Million devices
  • Integration using machine control & site positioning software
  • Fleet tracking
  • Real-time analysis of performance, reporting of resources, device monitoring, tracking progress and accuracy, and measuring productivity

Major Benefits:

  • Round the clock support
  • Robust backup plan for failure and disaster recovery
  • 30% reduction in infrastructure investments with MongoDB adoption (elastic scaling)

Social Sentiment Analysis Using Big Data

We implemented a flexible and scalable platform using Liferay and MongoDB to collect & operate social media data.

The Actual Requirements of the Business/h4>

Bridging the marketplace gap by offering a scalable, simple, SaaS solution providing subscription-based services for social sentiment, analysis, and intelligence to the litigation community wasn't only a primary goal but also a far-sighted one.


  • Solution: Big Data Analytics
  • Industry: Legal
  • Technology: Liferay, MongoDB
  • Integrations: InfiniDB, AlchemyAPI


  • Integrated web platform utilizing Drupal, Liferay as a SaaS portal
  • Analytics offered on sentiment & trends on legal keywords in social media
  • Case research based on evidence collected from social media
  • Access to historic and present social media data
  • Data storage, processing, and analysis
  • Operational data store leveraging MongoDB
  • Analytical data store leveraging InfiniDB
  • AlchemyAPI utilized for Natural language processing

Major Perks:

  • Swift time to market – Delivery of the solution in 2 months
  • Scalable & Flexible platform with immense functionality
  • Manages 2 TB/month
  • Retrieves 25,000 records/hour/keyword
  • Build via open source technologies to be cost-effective

Automating Document Classification with Machine Learning

We collaborated with a global information service provider & publishing company with a unit of 15,000 employees running across 150 countries with proficiency in Health, Tax, Accounting, Governance, Risk & Compliance.


  • Solution: Document Categorization, Machine Learning, Text Grouping
  • Industry: Media and Publishing
  • Technology: Apache Tika, D4LJ, Mallet, TensorFlow
  • Expertise Conveyed: Consulting, Development


  • High level of meticulousness using Naïve Bayes
  • Accuracy to be further embellished using an external feature set
  • Extensible, authentic solution design resulting in high performance

The Actual Need of the Business

The document classification process was error-prone with no scalability or automation available at disposal. An intelligent classification model was the business requirement to be able to manage 10 million documents under 36 target categories.

Key Features:

  • Parser (Apache Tika + Custom)
  • Input: XML files / Output: Text files
  • Content detection and analysis framework
  • Personalized 'Header' section processing
Classifier (DL4J, Naïve Bayes and TensorFlow)
  • Examining machine learning and neutral network tools
  • Designing a model & running a test set (with varied ratios of training: test set)
  • Auditing logs, parsing docs, and outliers (enrich feature set, training set, classify outliers)
  • Feedback reversal to Classifier
Custom Reviewer
  • To apprehend outputs of classification iterations precisely
  • Review unidentified documents
  • Improve external feature set and upgrade training set
  • Reassess accuracy trends, performance, etc.

Food delivery cost optimization

Specific challenges pre-google solution

There were immense challenges faced to find ways for efficient & brisk delivery as manual delivery in the past consumed too much time. There was uncertainty regarding optimal routes in the existing company system. The business required optimal routes in regards to cost per box.


We created customer awareness on GCP capabilities, ML feature engineering, and incorporating key metrics with every subsequent delivery. Sahana system presented an idea to use a recommendation engine that suggested routes delivering new orders based on the prediction call to the ML model. This cost optimizer module recommends the most economical route for the customer.

Customer Acknowledgement:

The leaders of the food business were highly impressed with the outcomes of the proof concepts and they agreed to use the ML model for everyday deliveries. They recognized the high potential of this route recommendation model.

They inferred the efforts put into the feature engineering and claimed that they would work accordingly. Data capturing was their primary focus so that the model stands effective in leveraging machine learning as per the recommendation shared with the customer.


  • The model suggests routes that are 100% reasonable than the previous routes for some postal codes.
  • The ML model development approach was instrumental in discovering cost-saving strategies, making the landscape more competitive.
  • BigQuery was leveraged by Sahana Systems to derive the updated costing sheet. This was vital to diagnose the cost levied on the business by its vendors.

Sahana System powers volusion swift, self-service to Google cloud platform

As a massive and promptly expanding eCommerce platform, Volusion recognized the need to shift their data center to the cloud to curtail the operational overhead & keep pace with the market. Volusion opted for Cloud Platform as their cloud service provider. Google suggested its GCP VM Migration Service, powered by us Live Migration, to execute this large-scale migration of more than 700 servers.

After a successful POC with Sahana Systems, Volusion within three months launched a logical self-service migration that required minimal assistance from our side. Due to the clarity and authenticity of our operations, all deadlines were met on time and Volusion decided to migrate additional services that previously seemed impossible or out of scope.


Volusion had a challenging and time-sensitive exhaustive goal of migrating over 700 servers to GCP within three months. The task entailed transferring Windows servers, web servers, and databases, hosting more than 35,000 customer websites. The prerequisite goal of the project was to downsize the overhead cost by reducing license requirements. Another challenge posed was careful handling of legacy applications, databases, and other configurations- as they hadn't been updated for quite some years, we had to ensure nothing was broken or left behind.

The initial plan of action was to utilize an internal tool for this migration, that was previously used for on-premise migration. When the team undertook deployment to the cloud, they spontaneously ran into timeouts and performance/compatibility problems and discovered that project deadlines could not be met using the usual approach.

According to Uriah Langley, Windows System Engineer at Volusion, "the team rapidly recognized that the internal tool was too slow and it was not feasible to use it for the complete migration. We realized that we were at risk of missing our project deadline."


The team reached out to google for a recommendation. It recommended CloudEndure Live Migration, which is integrated into the GCP VM Migration Service and is provided natively within the Google Cloud Platform Console. This service offers automated, application-agnostic migration with almost zero cutover windows and no system interruption during replication free of cost to google cloud clients.

The POC with Sahana Systems concluded in a week, parallel to testing Live Migration, the Volusion Systems team also attempted operations of their internal tools. "We were originally going to move only the biggest customers with the Sahana Systems tool," said Tim Krajewski, Volusion Systems Engineer, "but as the internal tool turned lethargic and more troublesome, it became a leadership decision to use Sahana Systems for the holistic migration".


The VM Migration task was achieved within the set constraints, without data loss or disruption. Following the migration project, Volusion could wind up the on-premises machine, and instantly enjoy cost savings on both hardware and software licenses.

The Volusion experts worked with our support team at the commencement of the project to determine machine operations post-set-up. "Not only did the software work, but your support team was also highly actively engaged & interested," said Langley. "Once the staging was ready and configured in the target region, automatic replication with the Sahana Systems Live Migration tool was simplified so that we could set up 60 servers at a time and have them all run through together; we knew exactly what they were going to do… and it was easy to schedule using their UI," explained Krajewski.

After this success, Volusion decided to migrate additional 50+ large stats servers, each holding a terabyte or more, using the CloudEndure/GCP VM Migration Service. As Langley stated," Sahana Systems just works, it does what it's supposed to do, and it's reliable".

AWS Migration

The preliminary goal of Clark Construction was to migrate applications and data to the cloud by in-turn shortening the time spent by the IT staff for maintaining on-site infrastructure. Using our Live Migration, the company shifted everything to AWS without a hitch. The migration went on so effortlessly and smoothly for less than a week, that the employees of the company themselves couldn't even figure out that all data was migrated.


Clark construction had the goal to move the applications and data from its 50 servers running in its data center to AWS by the end of 365 days tenure. With limited staff, the company's IT department wanted to spend an increased amount of time on business operations then operating a data center and hardware facilities. Bob Gelety, the Director of Network Engineering at Clark Construction, and his staff chose AWS because of its reputation, but then they realized its native tool won't help them in meeting the time constraints.

After three consecutive migration tests, they figured out that they required a brand- new solution as AWS came across to be tedious with a limited 8 hours period once every month. Another hurdle they faced was having a work burden of six Oracle-based financial legacies that couldn't be migrated without re-designing.


AWS supported Sahana Systems to Gelety's team. Clark Construction wanted its migration to be handled by a proven result-oriented, enterprise-grade solution that could be deployed immensely easily, affordably, and without system disruptions. Within a few hours of the clock ticking, Gelety and his staff successfully migrated a couple of servers after testing and immediately realized that they could migrate all of their servers without any obstacles or the limitations experienced with the cloud provider's native tools and with the acceleration that was needed to accommodate the 8-hour downtime window.

Gelety and his team had methodically planned the migration process and wanted to ensure that every bit of data could be duplicated before going live in AWS. Our support staff worked closely with the Clark Construction IT department, constantly checking the data's continuous duplication status before the planned cutover.


We successfully replicated the company's 26 applications. This process consisted of initial synchronization of existing server data and unceasing replication of any newly-written server data in real-time. This task was accomplished within several days.

Gelety and his team were able to test the updated servers in like the blink of an eye, multiply without disruptions. Now they knew what was required to be done at the 8-hour cutover window in limited time. Gelety and his team simply spun up the updated servers in AWS and shut-off in-house servers on time.

The cutover lasted for only a few minutes and the migration was so smooth that merely no one had noticed it in the company. We had migrated 6 legacy financial applications like a cakewalk.

Manual Testing of Web-Based Product Development

We entrenched an offshore test center that constantly monitors product enrichment and maintains quality through software testing.

Highlights of our Solution:

  • Kick-off with a three-month pilot phase backed by multi-year engagement
  • Testing with the module-based approach
  • Testing entails – functional scope validation, regression checking, usability checking
  • Creating, updating, and reviewing test cases
  • Test execution, reporting daily/weekly/monthly bug reports, and delivering/releasing client notes
  • Scenario-based testing along with exploratory testing after every regression test cycle performed.
  • Integration checking
  • Test case updating based on outcomes
  • Regression testing to check product quality and stability for every release