How to Deploy Upstream’s C4 Platform?




We made installation of the Upstream C4 platform into your production environment as seamless as possible. A lot of times our customers like to start with an offline dataset that they’re able to send to Upstream and then move from that into a production dataset they use as real-time data. Upstream solution is entirely cloud-based and it doesn’t require any agents to be installed inside the vehicles. The solution is also completely out of band, which means that there’s no inline elements that can create additional latency or bumps in the wire. The way that the C4 platform connects into your data feeds is by basically getting replicas of data that you’re already receiving from various sources. We tap into that data using two major methods. The first is using mirror ports, which basically does the mirroring at the network level. You can create mirror ports either on your network switch or on your cloud environment, and what it does is basically takes data traffic that you already have flowing into your cloud and creates a mirror of that and sends it over to the Upstream service.


The second methodology is to install open source Upstream collectors inside your production environment. Upstream’s collectors are open source agents that you can install either on your telematics servers or on your application servers. And then they take the data, serialize it from the servers into files, and transmit these files in a secure fashion to the Upstream cloud. Using the agents, we’re also able to compress the data and make sure that there’s buffering so that there is no data loss in the transfer from the collector to the Upstream service, and we’re sure that the full dataset is already received. The upstream service runs on a completely anonymized data. Which means that either the collectors themselves are able to anonymize the data before it’s being sent out or receive the data after it’s already been anonymized from a packet broker or from application servers that have this capability.


Anonymization typically means that data is stripped of any PII or Personally Identifiable Information and this can be various fields within a telematics dataset or a mobile or various application servers datasets. Once the data is received at the Upstream service, we have a collector module that is able to aggregate all this data from the various feeds, and then it also verifies that the data is indeed anonymized and there is no missing PII there. If it was able to identify that there is non-anonymized data, it’s able to create an incident right away then send that to the customer. An optional functionality that some of our customers use is elect to use our collector for its anonymization capabilities. The way that’s performed, the customer has to configure and provision which fields contain PII information, and that configuration is performed within the collector itself.


The collector is then able to parse the data before sending it out to Upstream and then any PII containing fields are then encrypted before being sent to the Upstream service. Only the customer owns the encryption and decryption key, and they’re the only ones that can actually decrypt the data for viewing and for monitoring. By using this nonintrusive approach, we’re able to seamlessly integrate within a customer environment, in some cases within a few hours, and then the upstream service itself can actually run within a customer VPC or in Upstream’s cloud giving you more options to run the Upstream C4 service based on your specific requirements.