Overcoming the Four Biggest Roadblocks to Cloud Bursting… with Ease!

By: Tom NiklJanuary 19, 2017

One of IT’s biggest burdens when building and managing a data center is how to optimize capacity. Essentially, any data center must be properly equipped to handle capacity at peak times, even though peak times may be infrequent. This can result in over-provisioning, with expensive computing hardware sitting idle for large periods of time. Some examples of “peak time” could be a retailer whose website will experience double traffic during the holiday season or banks doing year-end financial calculations. No matter the scenario, IT has been living in a world where their data center had to be built around peak- instead of average- usage cases. Which, of course, leads to higher hardware, maintenance, and management costs.

The idea of using the cloud to solve this peak time challenge has been floated frequently over the last couple of years. The general concept being that you build a data center to support average usage, and rely on the cloud for extra capacity when needed, either on-demand or during predictable seasons. Up until now, though, this has been easier said than done. There were four major challenges that IT had to overcome: data sets, application dependencies, cloud location, and performance. Let’s take a closer look at each of these issues and how Velostrata’s pioneering streaming-based workload mobility solution can solve them:

  1. Data sets: Most of today’s applications (and thus, users) rely on a large set of data to function properly. With traditional replication-based tools, this became a major challenge to leveraging the public cloud for bursting. It meant that for apps to work properly in the cloud, all the stateful data had to be copied, too. This would of course take hours (or days…), which meant significant time overhead before being operational. On top of that, once the first set of data was transferred, the data center and cloud would have to be re-synchronized to address changes that took place after the initial transfer. Then, the whole process had to happen in reverse when the cloud capacity was no longer needed. These data challenges were often enough to keep IT away from cloud bursting (and we haven’t even gotten to the other challenges yet!).

    Velostrata software gives IT the unique power to begin running their applications in the cloud on larger instance types within minutes while data remains on-premises. Any data changes made within the cloud instance are kept in-sync with the production data on-premises. This means that IT now has a way to add compute power within the cloud in minutes but without having to move all the data first.

  2. Application dependencies: In many cases, today’s applications are installed across multiple VMs and rely on other systems (like databases) to function properly. Because of these application dependencies, it wasn’t easy for IT to isolate an application and start it up in the public cloud when they needed extra capacity. It required too much manual labor to identify the dependencies and properly connect them within the cloud.

    Velostrata’s runbook automation alleviates these challenges by giving IT a way to properly plan and orchestrate the movement of multi-tier apps into the cloud. This means that IT can perform ‘lift and shift’ migrations quickly and with ease, regardless of their size. Even dozens (or hundreds!) of migrations are now possible, concurrently, within hours (instead of months).

  3. Cloud Location: Vendors like Amazon and Microsoft have cloud regions all over the world. This is great because it means a global IT can now provide low-latency access to any of their offices or customers. The problem, however, was that there was no easy way to take advantage of one or the other on the fly. Each region required some level of replication in getting the actual workload there so that when it was needed, it would be ready to go. That meant keeping multiple copies of the same app up-to-date across multiple cloud regions.

    Velostrata, however, decouples compute from storage, which means there is no pre-replication required for any region. Thus, IT has the power to migrate or clone any workload into any cloud region with just a few clicks. This gives IT the power to run workloads in whichever cloud region makes the most sense without having to lay hefty groundwork beforehand. Because IT can use any region, it also ensures that performance will remain high for all end users. Again, Velostrata handles the storage latency easily with advanced WAN optimization, deduplication and caching to ensure read/write performance.

  4. Performance: Companies quickly found that separating application components (for example: application servers in the cloud and databases centralized in the data center) introduced untenable latency. This made bursting virtually impossible for any application that required reasonable speed for data reads and writes.

    What’s truly unique about Velostrata is our ability to decouple compute from storage effectively. Patent-pending WAN optimizations, deduplication, multi-tier caching, and intelligent on-prem write-back make true hybrid cloud computing possible. All of this gives IT the power to quickly and easily move databases to the cloud along with the applications themselves.

These four challenges have been major roadblocks to IT organizations taking advantage of cloud capacity on demand (aka cloud bursting). But, with Velostrata’s streaming-based workload mobility solution, it’s possible (and incredibly easy) for any IT organization to move even stateful workloads into the cloud within a matter of minutes- any time they need the extra capacity. Whether it’s on-demand capacity  to address dev/test or unplanned projects, or planned peak capacity during a holiday season, Velostrata provides a simple, efficient way for IT to cloud burst while crucial data remains on-premises.

Want to see Velostrata’s cloud bursting in action? This quick 90-video shows you how to move a 1.5 TB workload that is struggling from a performance bottleneck in to Azure (where additional compute resolves the issue) in just five minutes.

Tom Nikl
Tom Nikl
Tom has spent twelve years leading product management and product marketing at technology companies large and small who focus on virtualization and cloud technologies. He currently blogs primarily about cloud migration, with an emphasis on overcoming challenges that companies face getting to the cloud and how to solve them. Prior to enterprise, Tom received a B.S. in Computer Science from San Jose State University. Outside of work he is an unabashed fan of Disney Theme Parks and various junk food. Find Tom on Twitter, too: @Tom_Nikl