Hadoop without HDFS
Discover the advantages of disaggregating compute and storage platforms.
The Hadoop framework is the most widely used among organizations that deploy Big Data clusters to handle large amounts of unstructured data.
But the different modules of Hadoop are now competing with more powerful solutions. Many of these substitute MapReduce with Spark for processing. And HDFS (Hadoop's distributed file system) is nearing its limits as more and more data is channeled into ever larger datalakes. Criteo, one of the biggest Hadoop users in the world, knows this.
There are few storage solutions capable of absorbing this type of load and serving the data to increasingly demanding computing clusters. But high-performance Object Storage technologies like the one developed by OpenIO can already replace HDFS.
Discover the benefits of disaggregating compute and storage platforms. Listen to those who have taken the step of successfully combining OpenIO and Spark.