In this special guest feature, Marcel Hergaarden, senior manager for product marketing at Red Hat, explains why he believes on-premises object-based storage is the correct approach for organizations that want better control over their data and greater cost savings. Marcel has deep expertise in storage and data management solutions, spanning technical sales and engineering roles. He is based in Amsterdam.
Although the public cloud offers great scalability and flexibility, there are many benefits to keeping data on-premises. For one thing, the egress costs of extracting large amounts of data from the public cloud can prove to be prohibitively expensive for many companies. And keeping data on-premises enables customers to control their data, relieving potential security concerns.
Indeed, maintaining data on-premises is particularly ideal for this moment in time. The pandemic has led to an inordinate increase in the amount of data being used. This is evidenced by an OpenVault report showing a rise in the number of users consuming 1 TB or more of data per month, with the average data consumed in the third quarter of 2020 up nearly 40% over 2019.
It’s easier to meet this demand when latency is minimized. That’s easier to do when data is stored on-premises and does not have to be retrieved from the public cloud.
And yet, on-premises data services management can present its own set of challenges, particularly if data is being handled in the traditional file storage manner. While file storage systems can store just about anything, they are limited in the amount of information they can hold. Organizations may need to add new systems to meet expanding demand, which can get very expensive and difficult to maintain. And since file systems only store a limited amount of metadata, it can be difficult for applications to find the data they need to run effectively.
This is where object-based storage becomes useful. Object storage uses a flat structure in which files are encapsulated into objects. Each object has unique metadata attached to it and can have additional labels. All objects are stored in object buckets. It’s an ideal and modern approach for organizations that wish to manage data services in-house. Here are four reasons why.
Infinite scalability
Since object-storage uses a bucket structure, rather than a file hierarchy, it is infinitely scalable. In fact, recent tests commissioned by Red Hat showed that object storage developed under the Ceph open source project could handle 10 billion objects — a number that testers expect to grow over the next few years. At the same time, object storage is cost-effective; when companies do need to add storage, they can easily do so as necessary without having to invest upfront in additional storage they may not need.
Better Searchability
Attaching unique metadata to each object makes for better searchability. There’s no need for file system crawls, which can be time consuming, especially when combing through large data sets. Applications can more easily and quickly find the data they need, resulting in faster response times.
Greater Intelligence
Leveraging AI and machine learning algorithms can give organizations actionable insights into usage patterns. Many companies use data lakes to store large repositories of information that can be fed into these AI and machine learning algorithms. Using object storage for data lakes enables organizations to make use of massive data sets that can be queried and used for highly demanding AI and machine learning workloads.
More Control
Employing on-premises object storage can help organizations assert control over their data by offering them the ability to implement unique bucket policies for data objects. For instance, objects placed in a specific bucket could automatically be replicated, while those placed in another bucket could be subject to data masking, providing a greater level of protection for highly sensitive or personally identifiable data.
Put all of this in perspective for businesses today. Think of a cable TV operator delivering video content bit-by-bit, instantly and on-demand for the customer watching television at home. Or, an online file storage system that must scale to accommodate the needs of thousands of remote workers around the globe.
These are the use cases that object storage was built to handle in a cost-effective manner. That’s why object storage is the most flexible and scalable data management option for our data hungry world.
Sign up for the free insideAI News newsletter.
