TrackVia’s Data Center Infrastructure Evolution: Part 1 of 3

trackvia_data_center_infrastructure

TrackVia’s Evolution in our Data Center Infrastructure:
A 2006 Start With Managed Services

Hello, and welcome to a new part of TrackVia’s blog! I’m Marc Haverland, CTO here at TrackVia, and I will be providing some insights into our engineering and production operations activities. We have a very talented group of engineers who have rapidly advanced our technology in the past year. With help from this stellar team, I hope to shine a light on some of the technology decisions we’ve made in building a highly available application platform, and some of the key tools, technology, and architecture approaches we’ve found to be successful.

We recently released a brand new product here at TrackVia: TrackVia Express, a lean, easy-to-use platform that enables customers to build their own applications and provide native mobile apps to their workforce for field data entry. You can check it out here. In addition to the fantastic user experience, the new product represents a significant evolution in how we view our data center infrastructure. This article, and the next two in this series, will walk through the key points along our evolutionary path, and will cover traditional managed services and colocation, a hybrid co-lo and cloud model, and a pure cloud model.

Our enterprise product has operated in a managed service model since its inception. In this model we have dedicated servers, and responsibility to manage everything above the OS. Our data center provider retains responsibility for the hardware, base OS, and remote-hands work in the data center. This model has offered both benefits and drawbacks.

Some of the benefits include:

  • A lower staffing level for Production Operations, as we can leverage the 24/7 on-site team at our provider; and
  • Reduced up-front capital expenses, as we lease the hardware instead of purchasing it.

Some of the drawbacks include:

  • A higher long-term cost, as you’re amortizing both the hardware and data center personnel expenses over time, with a vendor’s profit margin included in both;
  • More parties involved when troubleshooting hardware-network-OS issues;
  • Potential time-to-delivery delays because launching new services may require additional levels of participants: our managed services provider must purchase and install hardware from a vendor, and must provision the OS and connectivity, before we can begin deploying our new services.

When TrackVia was founded in 2006, operating on a shoestring budget, the benefits outweighed the drawbacks — a managed service model was the right way to go. (That calculus has changed in recent years, however, as public cloud providers such as Amazon Web Services, Rackspace Cloud, and others have matured. A different conclusion may well be reached today, and indeed often is, for early-stage startups.)

As TrackVia grew, a logical next step might have been a traditional colocation model, in which we take on ownership and responsibility for the hardware and network (presumably because the CapEx vs. OpEx tradeoffs show that such a switch is advantageous). Instead, however, we chose to pursue a hybrid, partial cloud solution.

In the next article we’ll provide an overview of this hybrid data center solution, some of the reasons why we chose this path, and how we’ve applied a hybrid cloud model with good results.