A survey of 656 IT leaders finds 70% of the data residing in the cloud is now stored in an object storage system.
Conducted by the market research firm UserEvidence on behalf of MinIO, a provider of an object storage platform, survey respondents expect object storage systems to store 75% of their cloud data in the next two years.
The survey finds the top applications making use of object storage include data analytics (54%), AI model training and inference (51%) and data lakehouses (44%). A full 62% of respondents plan to build a data lake using object storage in the next 12 months, with 30% planning to do so in the next six months.
The top drivers of further adoption of object storage is the rise of artificial intelligence (AI) at 52%, followed by increased performance (49%) and scalability requirements (44%), the survey finds.
Jonathan Symonds, chief marketing officer for MinIO, said that high ranking for performance helps dispel any remaining myths about the inability of object storage to meet the performance requirements of modern applications. In fact, a full 43% of respondents said that the typical network speed for an object storage cluster at their organization is 100GbE or faster. Additionally, many organizations are taking advantage of the latest generation of NVM-based platforms to optimize performance, he noted.
In the meantime, there are object storage challenges that IT teams need to address, including identity and access management (47%), data lifecycle management (46%) and encryption (44%), the survey finds.
Overall, the survey also finds that 64% of respondents work for organizations that are either all based in the cloud (19%) or mostly (45%). Many organizations continue to store large amounts of data using block storage systems and file servers in on-premises IT environments, but as more data is stored in the cloud it’s only a matter of time before enterprise IT organizations embrace object storage more widely, said Symonds.
In effect, many organizations want to be able to access a private cloud in an on-premises environment to run, for example, AI workloads in a way that has all the same capabilities of a public cloud, he added. That capability will also make it simpler to build and deploy applications spanning a hybrid cloud computing environment, noted Symonds.
There are, however, always storage challenges to overcome, especially when it comes to AI. Issues such as security and privacy concerns (44%), data governance (27%) and cloud-native storage (25%) top the list.
Additionally, the survey finds well over two thirds of respondents (68%) are concerned about the cost of running AI workloads.
The one thing that is certain is that the amount of data that organizations need to store is only going to continue to exponentially increase. In fact, the survey notes the top types of data being stored most often by organizations on object storages systems is application data (59%), followed by log data (45%), social media data (41%) and simulation data (40%).
The only way to address that requirement cost effectively is to rely more on object storage systems that are designed to scale. Otherwise, it will quickly become apparent that the cost of storing data will soon far exceed its value.
