Cloud Native Storage vs Third-Party Storage - Part 4
In Parts 1, 2 and 3 of this blog post we covered off; some great use cases for Cloud Native Storage solutions, we articulated that the solution you select should be driven by the needs of your application and covered off some of the leading third party cloud storage solutions in the market today.
In this Part 4 we will go through some of the design considerations you should think about when selecting your solution, whether it is Cloud native or a Third-Party solution.
Storage is hard
One of the key points to remember is that Storage is really hard, it doesn’t sound hard, but it is. There’s only 1 thing a storage system really needs to do, that is when I give it my data, it gives it back to me when, where and how I need it.
If a solution can’t do that then it’s not really a storage system.
There is a concern about native storage solutions that are trying to reinvent storage.
Doesn’t third party storage add a layer of latency between your data and the cloud native storage?
One of the major concerns about third party storage solutions is that Azure and AWS solutions should theoretically have the best performance as they will have no layer between application and storage.
Examples of these are AWS EBS and direct attached NVMe and SSD options.
This is not a valid point for the same reason that we don’t use bare hard drives and JBOD’s in the Data Centre.
We invest in Enterprise Storage for all the other important aspects and features it delivers.
We don’t just purchase hardware specs we buy a solution that actually works.
Common questions we then get are:
- Doesn’t proximity to the hardware matter? and
- Won’t having an abstraction layer result in lower performance?
The answer to this is “not really”, because raw hardware is for lack of a better technical term “dumb”.
As in do Enterprise storage arrays work because they have the 1 fastest drive NVMe or SSD drive?
No, they work because of things like RAID and the file system that has been built to read, write and optimise the Storage.
Solutions like NetApp Cloud Volumes and Pure Storage’s Cloud solution take the native cloud resources and do the RAID you need in the Public Cloud.
They stripe data across disks and instances, and they can mirror the data using erasure coding and they optimise the transfer protocols for you. The result is better overall, even if the abstraction means that data is further away.
Many times, these solutions perform better than native solutions.
Ease of Provisioning, Management and Scaling
Having a third-party storage array that works in the cloud is important but not as important as being as it being easy to; provision, scale, keep up to date, configure and manage as simply as you can with native cloud storage.
With solutions like NetApp Cloud Volumes and Pure Storage, you can provision storage on demand in a few clicks and you don’t have to build it, it just works.
The vendor manages the backend for you just like with the native cloud storage offerings.
In Part 5 the last part of this blog post we will discuss API’s and data governance and other considerations.
If you are interested in our other Cloud, Infrastructure and Security related Blogs please click here.