Tapad Engineering: Our Predictions for 2017

2016 was a busy year for us engineers at Tapad. We supported the launch of TV analytics as well as a brand new metric (VET), opened a data center in Singapore, and continue to improve our global real-time bidding system, which handles over a million queries per second with 99% being under 10 milliseconds. As we said -- busy -- but as we begin the New Year, there are a lot of interesting trends and developments that have us excited for the 2017:

1) Reasonably-sized-services

Microservice architecture has been a trend in 2016 and while we will continue to see this shift to lightweight services, enterprises will focus more on avoiding service proliferation and rather look at putting in place services that fit the context of the organization.

At Tapad, we see this as moving towards reasonably-sized-services centered around bounded contexts. We will continue breaking out large components into smaller services while also spending great efforts avoiding service-proliferation, but identifying the right “bounded-context” is what is key to this effort.

2) Machine-Learning-as-a-Service (MLaaS)

Machine learning infrastructure will continue to evolve. Tapad is active in this evolution as we continue to invest in our machine learning platform and automation to enable our Data Scientists to be as productive as they can be without having to worry about the technical infrastructure and environments. We also believe MLaaS enables new organizations to take advantage of machine learning without having to build out their own sophisticated infrastructure and data science teams.

3) Immutable containers

Docker containers are already the backbone of our infrastructure at Tapad, both for pre-production and production environments. However, utilization of containers will evolve even further in 2017. For example, at Tapad we are organically evolving towards immutable containers with our containerization initiative, where new Docker images are constructed each time application software changes, removing our reliance on separate configuration management tools.

Overall, this trend will be birthed by (1) the continued rise of DevOps (where we see software developers playing a larger role in operations and system administration) and (2) the wider adoption of functional programming (where immutable data structures reign supreme).

4) Start to the end of big companies

Realizing they cannot keep up with the innovation driven by small start-ups, big old legacy firms continue increasing their investments in startups in addition to experimenting with alternative organization structures in order to stay competitive. 2017 will be the year when they co-locate some groups at co-working spaces and set-up internal “start-up” units. We are already seeing this with traditional financial firms who are setting up internal start-up units in addition to invest in FinTech start-ups relevant to their business. Similarly, large Telecom firms are hiring space for small teams in co-working spaces where they sit side-by-side with entrepreneurs and startups.

Beyond 2017 and further into the future, a bold prediction is that these large organizations will continue to split up into many small separate units (bounded context=separate org unit) and eventually the small units will interact across different companies just like they interact with units in their own company.

What trends would you add to the list? Let us know or feel free to tweet us at @TapadEng.

If you’d like to join us as we continue to explore these trends in 2017, check out our current job opportunities here.