Search
Close this search box.

Azure Native Qumulo Now available in the EU, UK, and Canada – Learn More

Why Your Enterprise Infrastructure Should Innovate at the Speed of SaaS

Authored by:

How Qumulo relentlessly delivers innovation

At Qumulo, we keep our promises. One of the first promises we made, now proven by a decade of growth and nearly a thousand customers, was that we would ship code every two weeks and relentlessly improve our products.  

We made this promise because we saw that enterprise storage buyers were offered, at best, annual “major releases.” In enterprise data storage, vendors like Dell and Netapp resemble cruise liners, whose path is set early, changed rarely, and who move as fast as, well, cruise liners.

At Qumulo, we work iteratively, adjust to customer feedback, and rapidly deliver value to our customers. The first quarter of 2022 represented a particular moment of pride for me, as Qumulo’s VP of Product. Ten years into the company, the amount of progress we can make in such a short span of time continues to astound me. In reality, this was a relatively average quarter for us as an innovation delivery organization. 

In the first quarter of this year, we shipped seven software updates, 27 named feature releases and product improvements, and 24 API enhancements. The net effect of these improvements is extensive, ranging from making your data safer, your user experience simpler, and making it easier to move your data wherever you need it to be. My four favorite improvements deserve a little more air time: S3, last access time, instant upgrades, and write performance 

Serve All Your Unstructured Data

S3 is an increasingly popular protocol for unstructured data.  Developed by Amazon/AWS it has become the de facto object protocol and adopted by all major cloud and on-premises object storage providers. People are increasingly adopting S3 as an unstructured data protocol for use cases like internet storage, backup and disaster recovery, analytics, and archiving.  As people adopt S3, they expect to access their unstructured data using whatever protocol works best for their application.

Increasingly, customers are considering S3 as the 3rd protocol for unstructured data.  According to the Coldago end-users survey (2021), 25% of US respondents and 17% of EMEA respondents regarded S3 as their preferred access method to unstructured data.  64% of US and 65% of EMEA end users responded they need S3 and NAS access methods to the same content.

Supporting S3 as a first class protocol enables you to run “built for cloud” applications on premises for better performance, security or optimized cost. It allows you to support hybrid file and object workflows through the same storage system, simplifying workflows and avoiding unnecessary data movement. Furthermore, it enables you to run Apache Spark workloads directly against a Qumulo cluster using their commonly adopted S3 connectors.

One of my favorite elements of S3  is how well it plays with the rest of our protocols. At Qumulo, we help customers build massive unstructured data lakes, which can be accessed via SMB (2 and 3), NFS (3 and 4.1), REST, FTP, and now S3. As our VP of Engineering Molly Brown likes to say, “7 protocols, 1 data set.”

While we have only released our first slice of S3 support, it has already allowed us to receive crucial customer feedback to shape the direction of further developments and improvements. This is how we build products at Qumulo. We ship early and often, which allows us to iterate quickly and remove the guesswork from product definition. 

Identify and Archive Your Old Data

All storage systems were not created equal. If you need to identify cold data (e.g. data that has not been accessed for some time), and move that data to a more inexpensive storage option, then Qumulo’s implementation of last access time could be a huge asset. 

While we don’t claim to be the first storage vendor to develop a solution to this particular problem, the implementation of our feature is built to balance flexibility with performance, so our customers can get the data management capabilities they need without sacrificing performance in the process. Our straightforward implementation of a time comes with time granularity of hourly, daily, or weekly, and a cluster-wide setting with a default of ‘off’.  

The Best Upgrades in Unstructured Data Get Even Better 

We deliver new software relentlessly because it makes our products more valuable over time. However, to translate new code into value, customers need to upgrade. At Qumulo, we offer the simplest, fastest, and most reliable upgrade experience in unstructured data storage. 

At the heart of this delightful upgrade experience are a series of technologies that make upgrading Qumulo fast. This includes containerizing our software so that upgrading to a new version is nearly instantaneous, and an intuitive user interface to show you exactly where you are in the upgrade process. 

This quarter, we added a new “rolling” option for the infrequent occasions when an upgrade requires a reboot of the underlying platform (e.g. a firmware upgrade). This new experience keeps the cluster online and is non-disruptive for NFS clients. Most importantly, it is fast and simple to execute. 

Building great software is great, but we want to make sure any customer can take advantage of that new software easily. That’s why across a global, multi-exabyte fleet, more than two thirds of our customers are running software less than 6 months old. 

Faster Writes, Forever

It’s long been true that storage systems get slower over time, especially in their ability to rapidly ingest data. This is because of a fragmentation phenomenon in which data gets spread unevenly across a namespace, leading to complex calculations in the midst of a write operation. As a result, storage vendors either tell customers to expect their systems to slow down over time, or ask them to run costly defragmentation background jobs in order to reclaim performance. 

At Qumulo we set out to offer great long term performance without costly background jobs. To make that possible, we invented a new write allocation algorithm which places data in the optimal location across a cluster, and intelligently thinks ahead, in order to guarantee low latency writes no matter the age of the cluster. As a result, existing clusters can see as much as a 100% increase in write performance from this upgrade alone. 

The Roadmap is Included in Your Subscription

Qumulo’s world class engineers are maniacally focused on figuring out how to make our customers’ lives easier through our software. It’s an honor to work alongside some of the best minds in the world and watch how much they can deliver and how quickly they can deliver it through agile development practices and rapid product delivery.  

When you buy Qumulo, you’re buying what the product is capable of today, and tomorrow, and I can’t wait for what’s next.

Try Qumulo Today

Unlock up to a PB of free storage

What to Consider When Evaluating Enterprise Data Storage Solutions

Related Posts

Scroll to Top