Cloud-Ready, Software-Defined Storage, Flash Storage and More. Don’t Get Fooled by the Smoke and Mirrors
When it comes to storage, there are plenty of buzzwords being used within our industry these days. We see them in vendor ads, we read them in industry publications, and we sometimes even use them in our day-to-day jobs in IT. While they can be fun to say, these storage-isms can often cause confusion, lead to misinformed decision making, and even slow a business down. That’s why we decided to put some of the most common storage buzzwords through our critical lens to cut through the jargon and create some better working definitions for them.
Everyone says they’re in the cloud business – especially vendors who supply on-premises storage infrastructure. Whether they’re marketed as cloud-ready, cloud-enabled, or cloud-capable, their solutions don’t always fit into today’s cloud operating model. Some vendors that claim to be cloud-ready are actually only providing basic cloud features, such as copying system log data to the cloud, replicating data to a proprietary cloud platform for long-term storage, or simply putting analytical data in the cloud. The fact of the matter is, it takes a lot more than these fundamentals to accurately live up to the term.
To be sufficiently cloud-ready, the solution should be able to leverage public cloud resources to the max. This means that an array can tie into public cloud providers — such as AWS, Google, or Azure — to take full advantage of their flexibility and scalability in support of a hybrid architecture. A storage platform must also be able to replicate to your cloud platform of choice, support snapshotting and other advanced features within the cloud, and provide advanced cloud management and analytics capabilities, too. Otherwise, we feel the term cloud-ready is being used a bit too loosely.
Lots of storage vendors today say that they’re software-defined. While technically correct — without software you can’t do much of anything — the term is often used incorrectly. In some cases, vendors are touting software-defined solutions that can only run on their own hardware. As these systems age and need updating, a business has little choice but to purchase the next platform on the vendor’s roadmap. In this case, IT professionals who are looking at software-defined storage in an effort to avoid vendor lock-in might actually find themselves unable to break free.
When evaluating a solution as software-defined, it’s important to determine whether the system offers true hardware flexibility. This will be evident after a quick review of the list of platforms it supports. But be careful, too much flexibility can actually be a bad thing. When a vendor claims it supports every hardware option under the sun, getting quality support can become a challenge. So, look for a core set of hardware requirements that aligns with your hardware vendor of choice.
Many vendors have jumped on the flash bandwagon, hyping their solutions as the cure to today’s most common storage performance and efficiency issues. However, merely swapping spinning disk for flash storage doesn’t automatically create a better array. Sometimes the issues that occurred with HDDs will still be present on SSDs. This especially holds true when the array doesn’t intelligently manage data based on its performance and availability needs or when it uses legacy interface technologies.
In our opinion, using the term flash storage means surrounding SSDs with intelligent features and advanced technologies that make the most of their speed and agility. A good example is the intelligent use of media, where an array constantly analyzes data and assigns it in real time to flash storage, spinning disk, or RAM based on its needs. This can make storage a lot more cost effective to manage, without sacrificing performance. Another way to reap the rewards of flash storage is to connect it with an interface that is actually optimized for flash—non-volatile memory express (NVMe). By doing away with disk-era data access architectures, NVMe can significantly reduce power usage, increase IOPS, and drive down latency within an array.
Of course, there are countless other ways to improve the performance and efficiency of enterprise storage, and we’ll be writing about some of them shortly.
Ask anyone and they’ll tell you big data is a big deal. But if it’s so important, how can so many people have their own idea of what it actually means? Some use the term loosely to refer to all unstructured data, while others associate it with the contents of a particularly large storage volume. So it’s no wonder that when it comes to storage solutions, some vendors are riding the wave and describing their arrays as big data platforms.
Before we get into the hardware, what the term big data means to us is the entire process in which a business gathers, analyzes, and acts on data in an effort to make better strategic decisions.
When we talk about supporting that process with a storage array, one of the first things we look at is whether there’s sufficient processing power under the hood. Big data requires big performance. We also look for solutions that are actually cloud-ready (refer to buzzword #1), with the ability to easily move data to a public cloud service when additional horsepower is needed for an analytics job. The solution should also be certified and tightly integrated with commonly used applications in the analytics realm such as Spark, Splunk, and Hadoop. The way we think about it, a true big data platform needs to have performance, cloud integration, and tie-ins to all the analytical packages you need.
“Infrastructure as Code”
The reason we think there’s a haze around this term is that some vendors are well equipped to deliver it while others are not. Often times the greatest factor that separates the haves and the have nots is the timeframe in which an array was developed. Because Infrastructure as Code (IaC) is closely linked to the emergence of DevOps, which is relatively new, the younger arrays often have an edge. Meanwhile, some legacy storage vendors may claim to support IaC when they really only offer a basic API. From our standpoint, just because an array has an API doesn’t mean it delivers on the promises of IaC.
IaC represents a cloud-like operating model where you can provision resources quickly and within the natural course of your application development work. This augments DevOps environments with systems that make storage infrastructure way easier to manage than ever before. If you’re working with a vendor that has truly embraced the IaC movement, you’ll notice that their code is highly accessible and the resources you need are easily available to developers. This not only helps increase the speed of provisioning, it can also help improve the speed of your business.
Take the Doubt Out of Misunderstood Buzzwords
When used properly, buzzwords can help us remember complex topics and accelerate our understanding of a given concept. But when they’re ill-defined or watered down, they can stand in the way of what’s most important. This is true of cloud-ready, software-defined storage, flash storage, big data,andinfrastructure as code, which have begun to lose their meaning. By resetting your understanding of these buzzwords, we hope you can regain the focus you need to make the right decisions and capitalize on your next great storage opportunity.
To learn more about Veristor’s straight-talk approach to storage solutions, visit https://veristor.com/datacenter/enterprise-storage/cloud-storage/