Scale Up vs. Scale Out Architectures


Is Data Growth Extending Your Backup Window?

When looking to replace your tape library with a purpose-built disk backup appliance you may assume that all the big players in the market offer solutions that are similar in performance and capabilities. But the technical details, including the underlying architecture and how the deduplication operation is implemented, make a huge difference in the overall value of the solution. Let’s examine scale up vs. scale out architectures.

Choosing Between Scale Up vs. Scale Out Architectures

The most obvious reasons for choosing between scale up vs. scale out approaches are scalability and performance. Scale out architecture allows you to combine the power of multiple machines into a combined deduplicated data repository with each node having its own dedicated deduplication power. So, you are not limited to the server capacity of a single controller. In a scale up architecture scenario, however, you have a hard limit – the scale of the hardware on which you are currently running. Clearly then, one factor in choosing between scale up vs. scale out is whether or not you have enough resources within a single machine to meet your scalability requirements.

Solutions that have legacy front-end server architectures SCALE UP by adding resources to the single server. In this case disk shelves are added as data grows, but supported with no additional CPU or memory resources. As data grows, the backup windows keep expanding because of the fixed pool of available deduplication resources. Eventually, the backup window expands to a point where you must replace the front-end server with a more powerful server via a costly “forklift upgrade.”

ExaGrid offers the only disk-based backup appliances that offer GRID architecture that can SCALE OUT to respond to data growth. As a new appliance is installed, it gets added to the existing GRID. Each new appliance is a full server that adds more deduplication resources – processor, memory and bandwidth, along with its added disk capacity.

Scale Up vs Scale Out

Reasons for Choosing Scale Out Architecture Even If a Single Machine Meets Your Requirements

Today, with the availability of large multi-core and large memory systems, you might well have a single disk-based backup with deduplication appliance that can cover your scalability and performance goals. And yet, there are several other factors to consider carefully:

Continuous Availability/Redundancy – Having one big controller is a single point of failure. With a GRID, if one of the nodes fails the other nodes continue to function normally.

Cost/Performance Flexibility – As hardware tends to change quickly over time, you want to have the flexibility to choose the optimal configuration setup at any given time or opportunity to optimize cost/performance. If your system is designed for scale up only, then you are pretty much locked into the hardware that you are using.

Continuous Upgrades – Building a backup system as one big unit means that if a newer, more powerful controller is needed to keep up with expanding requirements over time, the only choice is to remove and replace the unit. Unexpected incompatibilities between older components and new appliances due to obsolescence may mean a complete forklift upgrade is needed.

More information on scale up vs. scale out architectures and how ExaGrid stacks up against other appliances is available at our Resource Center.

This entry was posted in Blog and tagged , , , , , . Bookmark the permalink.