February is the shortest month that usually ends unexpectedly, followed by the long-awaited spring. In the Northern Hemisphere it is the last winter Rubicon with cold and cloudy days when there are no clear signs but a strong feeling of the upcoming changes. Even birds are starting to chirp in February, telling everybody that very soon it will be warm and sunny. I do not know how it is in other countries, but in cold and wet Lithuania, where we have just around ten sunny days for the whole winter, spring always releases a tremendous amount of positive energy in people. We know and wait for it for the whole winter, like the Ukrainians are waiting for their upcoming victorious offensive that is also expected this spring.
As for flespi, we completed an important change in our network routing system that connects us with the Internet. The change happened on both logical and physical levels and was a prerequisite for starting a second datacenter and providing datacenter-level redundancy. Here is the scheme of how our network will operate in two data centers - QTS (current) and NorthC (new) with each hardware component replicated up to a PoP point in Amsterdam:
This change in the routing system to new wires and hardware was a reason for multiple downtimes last Monday, which gave us just 99.959% of total monthly uptime. All the information about how we handled the transition with the full history of messages is available in our NOC. We made a few human mistakes which could be avoided. And although the change itself is quite significant, we could do better in my opinion. However, now the routing system, which actual implementation together with a lot of new contracts with base layer suppliers took around 6 months, is now operating in a stable mode, and we are ready to implement a datacenter-level redundancy this year and provide uninterruptible services even in case our whole datacenter is unavailable for any reason.
- The biggest change of February is in our team. Finally, we were able to extend our engineering team with two high-level engineers with more than 10 years of combined experience in the telematics field - Yury Aheichyk and Evgeny Shatilo. Their previous job was devices integration in Wialon, and they were focusing on the advanced integration features. In our team, they will use previously gained experience and integrate video telematics to flespi according to our roadmap. If everything goes smooth with video, we will also consider tachographs, driver states, and .ddd files downloads via REST API as a next step in advanced telematics services flespi provides.
- Another portion of good news is that we increased the number of device groups available for each plan ten times as we noticed that they have become quite a popular building block for our users and initial limitations were too tight.
- We integrated two new device protocols into flespi: santana-rastreamento and x3tech.
- A set of changes was also in our analytics engine. We integrated the R-Tree index for all geofence testing calculations, which improved their speed by two orders in some cases. And we implemented a new messages counter in order to include source messages into the interval.
- For those who missed our announcement - keep a note that we are moving access tokens to another REST API path. Please update your integrations on time!
- Given the level of hype around generative AI, we also explored ChatGPT capabilities to generate correct answers in telematics.
- We made an overview of Suntech tracking devices that were most popular for us in 2022 as usual, amended with the priceless remarks by the Gurtam expert Vladimir Tihonchuk.
Last, but not least, significant event for the flespi team in February was the fourth week of the month when all of us, including Sergei Leuchanka who flew from Boston to Vilnius but initially landed in another country and city instead (like Biden), spoke about our shortcoming plans and clarified all steps on our roadmap implementation.
Here are the next steps we are going to take:
Within 1–2 months from now, you should expect from us a webhooks system that will be the core part of Automation.
Later this spring, we will create geojson-driven geofences as account-wide entities that will be available to use by reference in plugins and calculators.
In the summer, we will move external GIS services from plugins into account-wide entities, like modems. This will make it possible to configure Google Maps, HERE Maps, OSM, and other GIS services to be used for geocoding requests everywhere in plugins and calculators within an account.
With plugins, we will embed ML or algorithm-based filters that will be able to clean messages from devices from so-called “stars” - navigation jumps on a stationary device. Cleaned messages can be stored in flespi or pushed to a number of downstream integrations, which will eliminate the need for such functionality there as well.
Approximately in May, we will start the integration of video streaming devices and will introduce media hub functionality. I expect video telematics in its basic implementation to be available by the end of summer.
After that, we will investigate the possibilities of integrating tachographs and their .ddd files storage into flespi.
Vertical APIs for scooters, vehicles and their control methods are preliminarily scheduled for the second part of 2023 with detailed steps to be defined this summer.
A lot of work is scheduled for this year. This is together with the fact that we are growing at a 100% rate year-to-year with a lot of new clients and their projects that consume resources from our team and infrastructure. However, I’m quite confident that we will achieve everything we planned because we have a great team, we stay focused, and we know where we are heading with each particular step defined. And with such terms we cannot fail by definition!
The following picture clearly illustrates how much we are going to accomplish this year ;)