Why is it still rare to enjoy a stable internet connection onboard trains? What is it that makes it so hard to get right?
A train may host hundreds of passengers with WiFi-enabled devices, along with several other connected onboard systems, much like a large office or venue. However, unlike buildings, trains cannot rely on wired connections for obvious reasons. They must depend on wireless technologies, and when combined with the large number of users, this creates a unique niche use case.
Providing a high-capacity reliable internet connection onboard a fast-moving vehicle faces a unique set of challenges that stems from the fact that trains:
- travel at high speeds through remote and un-inhabited areas, often far from main roads
- host potentially hundreds of passengers that together generates thousands of simultaneous data sessions
Mobile network operators typically design their infrastructure to ensure robust coverage and high capacity in densely populated areas and along main roads. Trains, however, often pass through sparsely populated regions where the signal may be weak or even nonexistent.
To counteract weak signals, multiple mobile networks and satellite communications can be used simultaneously — a technique known as aggregation. The purpose of using an aggregation function on trains is to minimize communication disturbances while maximizing capacity. Using multiple communication links introduces a number of new challenges.
Key Challenges
- Transparency: Each data session is identified by its source and destination addresses. If either address changes during a session, the connection breaks and must be re-established. Since mobile or satellite network providers assign an address to each communication link, moving a data session from one link to another forces the client to re-initiate the communication, which leads to service interruptions. To address this, an aggregation function is split into two parts: one onboard and one onshore. These parts work together to dynamically move fragments or entire data sessions across the available communication links. This approach effectively preserves the integrity of the data sessions, making them appear to originate from a static address that remains constant throughout the session lifetime, even if a session is moved between links
- Reaction to Change: As the train moves through different surroundings, signal properties constantly fluctuate, resulting in highly dynamic communication link characteristics. To maintain optimal performance, it’s crucial to closely monitor these characteristics and adapt the utilization of available communication links accordingly
- Overhead: Enabling transparency through aggregation requires additional signaling in addition to the “useful” data being transferred. Since capacity is already a scarce resource on passenger trains, any extra signaling reduces the available capacity for useful data transfer. Therefore, the aggregation function must be carefully designed to minimize signaling overhead
- Variation in Link Properties: Trains use various network technologies, such as 4G, 5G and satellite links, which have inherently different properties in terms of delay, jitter, loss, throughput, and cost per unit of data transferred. The aggregation function must efficiently manage these differences in a robust and efficient manner
The Solution
At first glance, it may be tempting to base the aggregation function on existing bonding protocols, such as Link Aggregation Control Protocol (LACP). However, as mentioned earlier, the challenges presented in the train context are unique and fall outside the scope of those protocols. Drawing on 15 years of experience in train communications, Oxyfi has developed and refined a robust, high-capacity aggregation function specifically tailored to the demanding railway environment — making it one of the best aggregation technologies in the world today.
