This is a guest post written and contributed by Marc Haverland, CTO of TrackVia, a Rackspace Hybrid Cloud customer. The BIG Idea behind TrackVia is that average, non-technical business people should be able to design and build their own applications to better and more easily track, manage and do their work, whether it’s individual tasks, department programs or companywide functions.
TrackVia is all about empowering not-so-tech-savvy business users to build their own custom apps in an easy drag-and-drop model with zero coding. We were founded on the principle that applications are best designed by the people who will use them every day to get their jobs done.
We started out as a small bootstrap operation, but have started to scale up over the past two years as more customers have flocked to our service. We’ve grown to nearly 2,000 customers who have built several thousand different apps on our platform, including CRM, Marketing Automation, Support Ticketing, Product Management and other workflow or analytics applications tailored exactly to their needs.
Our infrastructure has to be stable and highly available—it has to just work. Many of our customers’ applications are mission-critical for them—they track their customers, orders and projects that drive their businesses and revenue streams. What good are we to our customers if our technology isn’t up to snuff?
We first came to Rackspace through our need to build out a thorough and complete disaster recovery (DR) plan to supplement our traditional managed service/colocation model. We stood up a full DR site with Rackspace and built it on the hybrid cloud model, which will enable us to grow as our business does. It also gives us the right cost-efficiencies that we need for DR purposes.
A hybrid cloud gives us the best of both worlds. While we have a goal to be pure cloud at some point, we have some elements that just don’t do well in the cloud in their current form, which is why we maintain dedicated hardware for those elements. Our large datastore tier currently requires dedicated hardware for performance reasons, while our SaaS web application and presentation environment operates very well on-demand in the cloud. Our DR costs are reduced by paying only for the dedicated hardware tier 24/7 (to which we’re constantly replicating live data), and we cost-effectively scale up the cloud-based tiers on-demand during a failover event. The net result is the right balance between a highly-available and performant DR capability on one hand, and a low-cost, on-demand cloud cost structure on the other hand.
The move to a hybrid cloud wasn’t one we took lightly. We did our research and our homework. We looked at Amazon Web Services, but AWS didn’t work out for this part of our service due to its lack of dedicated servers. (We continue use AWS for many things, but a cloud-only option wasn’t the best fit for this specific application.) We also considered SoftLayer, but its way of connecting dedicated and cloud environments was complex, and complexity can and did introduce higher risk and more difficult troubleshooting.
Currently, we run a number of dedicated servers as data replication targets from our primary data center, and through RackConnect those servers seamlessly coordinate with Cloud Servers on which our application and presentation tiers run, all on the same private networks. We also leverage Cloud Load Balancers in front of the overall stack.
We are rapidly progressing toward moving our entire infrastructure to the cloud, but until then, we need the stability and performance of dedicated servers coupled with the flexibility and on-demand cost efficiencies of the cloud servers; so a hybrid cloud was the obvious choice, and one that has worked well for us.