TerramEarth manufacturers heavy equipment for the mining and agricultural industries. About 80 percent of their businesses are for mining and 20 percent from agriculture. They currently have over 500 dealers and service centers in 100 countries. Their mission is to build products that make their customers more productive. The workload is divided between nine terabytes per day delivered streaming over cell service. That's the IoT part, and 900 terabytes per day delivered via G zip CSV file. Currently, the G zip CSV data from the field takes about three weeks to make it into the data warehouse. That means some customers have vehicles out of service for four weeks waiting for parts. The company knows that IoT is coming and is preparing to meet the changes as traffic shifts from the file transfer model to the cell IoT model. However, the data warehouse is behind technically and also is not meeting customer business needs. They need a data warehouse upgrade that will handle both problems. The company provides heavy equipment through 500 dealers in 100 countries. One of the company goals is future-proofing, being prepared with skills and infrastructure to keep pace with changes in the industry. Immediate business goals are to decrease downtime. Downtime can currently run beyond four weeks. They would prefer to get this down to about one week with an average being between one we can four weeks. They want to give those 500 dealers greater access to customer data. For example, if the dealer knew that a particular machine was out of service often, they might want to approach a customer with an upgrade offer. The 20 percent of the business in agriculture vertical is a potential growth area. The company wants to be able to partner with others net industry to expand their market. One way to do this is to be in the lead technologically. So, when there are inflection points in the market, TerramEarth can take advantage and enter new niches. This says it's about vehicles, but we can look at it as a way to understand the data demands it will be placed on the IT infrastructure. They have 200 thousand vehicles connected to sell networks that stream data at the rate of a 120 fields per second. They're on for 22 hours a day resulting in nine terabytes of data per day. They also have another 20 million vehicles that are not cell connected. The data is stored on the vehicle and downloaded when the vehicle is serviced. It takes quite a bit of time for the data to be uploaded and transformed. Currently, that process causes about a three week delay. To cover any parts that might be needed over a three week period, TerramEarth is forced to keep additional parts in stock. If the feedback were faster, their spare parts could more closely match the real demand and they could liberate some of the money that's currently invested in stock and sitting on shelves. A few key points. The first one is that this is not an ERP project. ERP is electronic resource planning. If you instrument your parts demand through your business and map it back to the supply chain, you can do things like order parts automatically when they're needed or even look at suppliers inventory and preorder parts if they are not enough in stock. In any case, TerramEarth has decided that ERP should be an upgrade for another day. They'll stick with current surplus supply levels and maintain the same supply chain as today. They're happy just to provide visibility to supply partners. Change is already occurring in the management team through retraining management about emerging opportunities. TerramEarth realizes that the technical staff will need to be trained on the new systems and solutions. Moving the data warehouse to BigQuery will handle a lot of the main customer and business issues having to do with parts delay. It will need a front end that can handle today's IoT demands and will grow and adapt to the changing categories of demand as more streaming solutions are employed and fewer file-based solutions. Keep in mind that Cloud IoT core doesn't suffice to get your data to cloud storage. Cloud IoT core brokers between IoT devices and Cloud Pub Sub. You'll almost certainly want Cloud Dataflow to get the data to the next place. For vehicles that store and upload their data, the service will need to handle 900 terabytes of batch data transfer each day. Go ahead and define your solution. Consider how you would design this solution for a real customer. What are the questions you want to ask to help clarify your design? When you're ready, we'll look at one sample solution
