メインコンテンツへスキップ

What Customers Are Saying About AI

Mike McNamara
Mike McNamara

The artificial intelligence (AI) survey data discussed in the blog post is based on answers from more than 100 customers at recent AI conferences to gather primary research on AI data management challenges, AI infrastracture and tools, and reasons for choosing different types of data storage.

The majority of survey respondents were software developers (24%), data scientists (18%), researchers (15%), line-of-business owners (7%), data analysts (6%), or general IT (6%). Company size of respondents ranged widely: more than 10,000 employees (45%), fewer than 1,000 employees (28%), 1,000 to 5,000 employees (18%), and 5,000 to 10,000 employees (9%). The top use cases for AI were healthcare and computer-assisted diagnosis, automotive and autonomous vehicles, manufacturing and robotics, and financial services and fraud detection.

The top three storage and data management challenges with AI were scaling storage, cloud integration, and backing up data.

What Customers Are Saying About AI - Inline Image 1The top three requirements when choosing an AI infrastructure vendor were cost, easy to deploy and manage, and services and support offerings.What Customers Are Saying About AI - Inline Image 2The three most common tools used with AI were NoSQL databases, Apache Hadoop, and Splunk. The majority of respondents either use a Hadoop cluster (data lake) with AI or plan to in the near future.

The most common file systems used with AI are HDFS and NFS. Less than 10% of the respondents used ZFS and GPFS.

Cloud storage is the most popular storage used for AI, followed by direct-attached storage (DAS) and then external storage.
What Customers Are Saying About AI - Inline Image 3The top reasons for using servers with internal storage or servers with JBOD are performance, cost, and management decision. As shown in the following graph, the top reasons for using external data storage are performance, reliability, and data protection.What Customers Are Saying About AI - Inline Image 4NFS is the protocol of choice for AI with external storage, followed by Fibre Channel and then NFS. Ease of use, easy to scale, cost, and already use cloud for compute are the main reasons for using cloud storage. The most popular services used in the public cloud are Amazon Web Services, Microsoft Azure HDInsight, and Google Cloud Dataproc.What Customers Are Saying About AI - Inline Image 5To learn more about AI and how it’s transforming how business processes are carried out in the digital era, read the “Infrastructure Considerations for AI Data Pipelines” report from IDC. To learn about NetApp® AI solutions, visit www.netapp.com/ai.

Mike McNamara

Mike McNamara

Mike McNamaraは、NetAppの製品およびソリューション マーケティング担当シニア リーダーであり、25年以上にわたってデータ管理とクラウド ストレージ マーケティングに携わってきました。10年以上前にNetAppに入社する前は、Adaptec、Dell EMC、HPEで勤務していました。また、主要なチーム リーダーとして、ファーストパーティのクラウド ストレージ サービスや、業界初のクラウド対応AI/MLソリューション(NetApp)、ユニファイド スケールアウトおよびハイブリッド クラウド ストレージ システムおよびソフトウェア(NetApp)、iSCSIおよびSASストレージ システムおよびソフトウェア(Adaptec)、ファイバチャネル ストレージ システム(EMC CLARiX)の発売を推進しました。過去には、Fibre Channel Industry Associationのマーケティング分野の議長を務めたこともあり、Ethernet Technology Summitの諮問会議や、Ethernet Allianceの現役メンバーとして、業界誌に頻繁に寄稿しているほか、各種イベントにスピーカーとして数多く登壇しています。さらに、FriesenPressより『Scale-Out Storage - The Next Frontier in Enterprise Data Management』というタイトルの書籍を発行しているほか、KaposによりB2B製品マーケティング担当トップ50に選出されたこともあります。Mike McNamaraのすべての投稿を見る

次のステップ

What Customers Are Saying About AI | NetApp Blog