Case Study 1: Scaling Cloud Infrastructure
Scaling and streamlining complex cloud infrastrucure, faster and at a greater scale
A Data Reseller approached Foretheta to modernize existing cloud infrastructure. The client wanted to reduce the cost of ownership of their existing infrastructure while having the flexibility to scale their data processing to higher traffic rate when needed.
The client was using Rackspace Dedicated Servers. Their workloads were low most of the time, so they were mostly wasting server capacity when it was unused. To meet the potential demand spikes, they would need to lease even more servers.
Foretheta helped migrate the client's infrastructure to AWS to enable them to autoscale and handle the increased workloads cost-effectively.
The client's had a redundant (primary/backup) set of servers to host their Celery workers. The celery workers pulled jobs off a Redis cluster. A redundant MySQL server set was used for persistence.
Foretheta replaced Redis with ElastiCache. MySQL servers were replaced with RDS. The Celery workers were migrated to AWS Lambda. Foretheta helped the client entirely move away from dedicated server capacity in Rackspace. The new technology stack could easily facilitate scaling out capacity based on demand. This lead to a 63% reduction in costs.
Case Study 2: Fast Text Analytics
Driving greater understanding through fast Text Analytics
A Patent Analytics company wanted to speed up their text analytics jobs that they ran on 2TB of compressed Patent text data in the cloud. The legacy application took days to run some of the data transformation operations on more than 100 Million documents. The client had provisioned fixed capacity to perform text analytics operations on the Patent data. Since fixed server capacity is expensive, they had to limit the size of the servers to fit their budget. The resource-intensive jobs utilized all of the available capacity very quickly. Subsequently, the jobs took a very long time to complete.
The slow jobs also meant that the client's developers were unable to iterate fast enough and think of new value-added innovations to apply on top of their Patent data. The client desperately needed to reduce the time required to run data transformation jobs on their data from days down to minutes, while remaining within budget.
Foretheta moved the client's infrastructure off of their bare-metal servers and onto Data pipelines that scale-out when needed. Also, this allowed the client to run jobs on the 100 million plus documents in parallel. Foretheta used Amazon EMR to make running Hadoop clusters smooth and fast on 100+ million documents. The changes allowed the client to reduce the time it took to run their jobs from days, down to less than an hour. Another advantage was that the client's team could now focus on innovating and creating more value-add services on top of the Patent data, instead of being bogged down by long-running jobs.
Case Study 3: Aligning Remote Workforce
Aligning remote workforce to work towards the right business goal
A major staffing agency approached Foretheta to help them align their workforce to their strategic plan. It was challenging to execute against their goals by tracking the performance of each employee against their KPIs. Tracking and execution had to happen in a single place and on a regular cadence. Their existing system depended on slow tools that did not integrate well.
The client wanted to build a custom application that would fit their workflow. They had less than a hundred users, so they intended to keep costs low as they ramped up their project. At the same time, they wanted the ability to scale, when needed. The client was based out of a smaller region and cared about the ability to find a technical team that can maintain their project as they found success.
Case Study 4: Testing Price Impact of Order Book Events
Validating algorithmic trading strategy
A client approached Foretheta to help validate an algorithmic trading strategy they were developing. The strategy involved testing the price impact of order book events. The client needed to implement and test the strategy in a short amount of time. The client required an option to take the trading strategy live if needed.
Speed was of the essence as the client was working under a strict deadline. The implementation needed to make use of existing systems that provided access to historical and live data. The client required a system that implements existing API for frequently used models in algorithmic trading strategies.
The client and Foretheta selected Lime Brokerage Studio as it had access to live and historical data on the platform. Foretheta quickly developed the trading strategy and validated it. The client had the option to run it on live data. The trading strategies were deployed to hosted machines with low-latency access to multiple US markets. They also had a rich set of features for backtesting intraday strategies.