This semiconductor and telecommunications equipment company designs and markets wireless telecommunications products and services. It derives most of its revenue from its chipmaking and patent licensing businesses. Based in the U.S., the company has more than 200 locations worldwide.
The company’s yield management team archives all semiconductor test data from test houses and foundries worldwide and makes that data available for analytics. The chip manufacturer required a solution to store, access, and process large amounts of data quickly. They also needed to be able to handle data from a wide range of sources in a multitude of formats. Their previous system was unreliable and slowed down their ability to report to customers.
"Our old system was a point of failure because it created a bottleneck in the import and export of data and did not have high availability features built into it," says a senior staff analyst at the company. "MapR gave us the features we wanted at a fraction of the cost."
The company chose the MapR Converged Data Platform (MCDP) because it gave them fast import/ export speeds, high availability, and NFS capabilities while reducing their infrastructure costs. They looked at Cloudera and Hortonworks, but only MapR gave them all the features they wanted at the speed they required. In addition, the MCDP provides easy backups with Snapshots and additional analytics features at one-third the cost of their former system. "Because MapR works with Hadoop, we get an entire ecosystem with machine learning tools and packages," says a senior software development manager.
The chip manufacturer is also using NFS capabilities to allow faster access to their data. "By using MapR and its excellent NFS, we are able to give our customers access to the data directly from the source through NFS for machine-to-machine programmatic access," says the senior software development manager
This chip manufacturer takes advantage of MCDP high availability capabilities including snapshots and mirroring.
"Our old system cost three times more than MapR and to make it high availability, we would have had to double our storage requirements, effectively making it six times the cost of MapR."
-- A Senior Staff Analyst Leading Chip Manufacturer
Performance was one of the main reasons the chip manufacturer chose the MapR Converged Data Platform. The company’s goal is to process, archive, and make their data available to end users as fast as possible. "Our system is doing a lot of validation, enrichment, and archiving of data but the end goal is to get the data to our customers. In some cases there’s a financial impact if we don’t move quickly enough," says the senior software development manager.
"Using MapR, we achieved import and export speeds that were ten times faster than our old system," says the senior staff analyst. The chip manufacturer is seeing real benefits from this increased speed. "Test time costs money so if data is available quickly, engineers can make faster decisions that can immediately be fed back to the test floor, which saves us money."
The chip manufacturer estimates they are able to run MapR on commodity hardware at one-third the cost of their previous system. When they added high availability, their savings doubled. "Our old system cost three times more than MapR and to make it high availability, we would have had to double our storage requirements, effectively making it six times the cost of MapR," says the senior staff analyst. "The cost savings with MapR was huge and we’re getting high availability."
MapR has proven to be extremely robust in production. "We went live more than two years ago. In that time, we have had zero downtime due to MapR," says the senior software development manager
"We went live more than two years ago. in that time, we have had zero downtime due to MapR."
-- Senior Software Development Manager Leading Chip Manufacturer