Business Intelligence Companies & Consultants in Eugene, OR

Moonshadow Mobile, Inc. is a software engineering company based in Eugene, Oregon. It supports organizations whose data streams have become too large to move, store, search and analyze comfortably. Rather than recommending more servers as data volumes climb, the firm redesigns storage formats and processing pipelines so that the hardware a client already owns can carry a far heavier load. It states that its patented approaches frequently make pipelines ten times faster and lower cloud costs by 80 percent.

Who They Serve

Moonshadow's work suits organizations that generate large volumes of structured records. The company points to a range of sources and fields:

• Mobile apps and IoT devices
• Connected vehicles and traffic analysis
• Cloud servers and server farm statistics
• Time-series data, including weather and stock market information
• Population databases such as voter registration files

A showcase example is the California registered voter database belonging to its customer L2 Political. It holds roughly 16 million records across 649 columns, which adds up to more than 10 billion values. In Microsoft SQL Server the database occupies about 92 GB, while Moonshadow's Ephemeris datastore holds it in about 8.9 GB, an average below one byte per value.

Services and Approach

The firm offers two routes. Its engineers can build custom data formats and complete processing pipelines, or they can target the one stage that acts as a bottleneck. Alternatively, clients can integrate Moonshadow's tools into their own environments. The company says trillions of records have passed through pipelines its engineers created.

Each project begins with an analysis of the existing storage formats and data streams. From there the team works at the bit level instead of the byte level, which removes the unused space found in most formats. It also trims excess precision.

Latitude and longitude values, for instance, are often supplied to a centimeter even though the GPS devices that produced them are accurate only to about five meters. For time-series data, the company notes that compression can improve as the volume grows, so a tenfold increase in data may produce only a doubling in size.

Products

MMZIP is a lossless compression tool for columnar data that customers run on their own servers, behind their own firewalls. It comes as a command-line tool or a C library and works on standard x64 processors. It examines an initial sample of rows to choose a compression strategy and adjusts if the data changes shape.

Its streaming design lets a terabyte-scale file be handled on a machine with 32 GB of RAM, and compression, transfer and decompression can overlap. During decompression it can filter records into smaller project datasets, derive date and time fields, add place fields, assign road segments or zones to waypoints, and calculate differences between points such as acceleration, time and distance.

Ephemeris is a patented, software-only NoSQL in-memory datastore for ordinary commodity servers. It packs values at the bit level and keeps recurring combinations of related fields, such as ZIP code, city, county and state, as single entries. Its on-disk layout nearly matches its in-memory layout, so loading is close to a straight copy. The company cites access times near 10 nanoseconds per record and says some customers have cut their server counts by hundreds.

Big Data Visualizer is a web-based mapping tool built around the familiar Bing Maps environment. Users filter and color-code geospatial data by any combination of attributes, switch between map, satellite and hybrid views, and see counts and percentages refresh as they zoom or pan. Nothing needs to be installed, and it runs in a browser on Windows or Mac computers.

What Sets Them Apart

Moonshadow's central idea is efficiency over expansion. By shrinking data, speeding retrieval and presenting results on maps, it aims to cut storage and transfer expenses while keeping data available for analysis and business intelligence. That focus may appeal to teams whose geospatial, vehicle or time-series workloads are outgrowing their budgets.