Scale out storage systems deliver the ability to start small and grow as the business grows and the data volumes increase. This makes scalability a very important ability for data center infrastructures; especially for data driven organizations and companies.
But what do we really mean by scale out storage systems? And why do we say that it goes a long way? We’ll be answering these questions in this article.
What is Scale Out Storage?
Scale Out is the ability to add more storage capacity to your storage systems when you need them but it’s not just that. There’s more to it.
Actually, scale out and scale up are sometimes mistakenly equated as the same thing but they’re really not the same. Scale up is the addition of storage capacity to existing data storage systems. However, scale out is the addition of not just storage capacity but also performance.
Scale up is also called vertical scaling whereas scale out is referred to as horizontal scaling. That just adds to the confusion, doesn’t it? Let’s take care of that then.
How Scale Up Works?
To scale up, data center administrators integrate JBODs (Just a Bunch of Drives) or EBODs (Expandable Bunch of Drives) with their storage systems.
As the name implies, JBODs or EBODs, are only a bunch of storage drives. They do not have any processor or motherboard that can handle the processing end of the storage. They only add more storage space to the existing storage system.
If you’re thinking: then that just adds more workloads. Isn’t that going to be slower?
That’s a good question and you’re spot on. If the processing capabilities of the existing storage system aren’t capable enough, then scaling up doesn’t end up being the good idea. That implies that there’s a limit to how much you can scale up. After that limit, you can add more storage space but it’ll just slow you and your organizational processes down.
How Scale Out Works?
To scale out, data center administrators have to integrate complete servers with the existing storage system.
That means the server has everything in it including the storage drives: the storage controller and the RAID controller (if applicable). Thereby implying that with the addition of a server, the data center administrator adds the ability to compute too because each server has its own processor.
As opposed to scale up, this actually improves the performance of storage systems. How?
Let’s say there’s a single data storage server, and then another is added to it. Now, there’s 2x processing capability along with the 2x storage capacity. The totally workload is aggregated over the two processors. Therefore, there’s an incremental increase in processing and performance.
Unlike scale up, which is limited by the capabilities of the processor, scale out doesn’t have any limits. Good storage vendors actually offer the ability to scale out to virtually unlimited number of storage appliance nodes. That’s why we say that scale out goes a long way, literally.
Still confused about scale up and scale out? We recommend reading this blog: Understanding Scalability in Data Storage
Now that we’ve established what scalability is and what we mean by scale out, let’s address the question: how does scale out go a long way figuratively?
Scale-Out Storage Systems “ They’re Future-Proof
There are two reasons why we recommend scale out storage systems; excluding the obvious increase in performance:
- Future-Proof Storage Solution
- Cost-Effective Storage System
How are scale out storage systems future-proof you ask? Well, that’s a rather simple one.
Scale out storage systems are highly scalable. Data growth isn’t a matter of if for digitally transformed work environments, it’s a matter of when . And scale out systems can scale whenever your data grows more than the storage capacities of your on-premises storage systems. That’s what makes it a future-proof solution.
When compared to legacy storage systems that only increase in complexity when scaling, purpose-built scale out storage systems are truly a future-proof solution.
Okay, so the future-proof part makes sense but how are scale out systems cost-effective?
Scale out storage systems are a good decision from a TCO (Total Cost of Ownership) and ROI (Return on Investment) point of view. How?
With legacy infrastructures, businesses have to thick provision or fat provision. Thick provisioning or fat provisioning is usually related to virtual storage resources but we’re using it here in a similar context. In anticipation of future growth, businesses purchase more storage servers than they need immediately. The main driver for this is to prevent inconvenience later. That surely seems like a good idea but it incurs larger CapEx. And consequently, the TCO doesn’t look good.
Scale out storage systems on the other hand, allow companies to start small and then grow as the data grows. When CFOs translate this ability into numbers, they realize that it’s effectively reducing the initial CapEx by half; similarly, the TCO is improved too.
Before we conclude this article, this needs to be mentioned that the ability to scale out isn’t just related to storage systems such as NAS or Unified storage systems (NAS + SAN + Object), and Hyperconverged Infrastructures (HCI). It’s also very important for purpose-built backup and disaster recovery solutions as well.
Conclusion
The only guarantee about enterprise data is that it will continue to grow. The continuous growth of data demands that if you’re investing in an on-premises data center then it’s wiser to invest in a highly scalable and scale out storage system.
With the scale out systems, your business is prepared for the future data growth and when you calculate the ROI, it comes out to be far more economical then initially inexpensive consumer-level storage systems.
Enterprise use-cases demand enterprise-grade systems and scale out storage systems are definitely one of them.
