Amazon SageMaker, Amazon Web Services’ platform dedicated to cloud-based machine learning, has announced a series of significant updates for the year 2023. These innovations are designed to improve both collaboration and disaster recovery, key aspects for managing and supporting critical data in machine learning projects.
One of the highlighted implementations is the enhanced release of SageMaker Studio, which now incorporates new applications like JupyterLab and Code Editor. Each application has its own Amazon Elastic Block Store (EBS) storage volume, allowing for more flexible and efficient storage management. Additionally, custom instances integration of Amazon Elastic File System (EFS) has been introduced. This facilitates file and resource management in custom environments, catering to the specific needs of users.
SageMaker has strengthened its focus on disaster recovery, crucial for users who rely on the service for critical tasks. Using Amazon EFS replication capability, the platform ensures operational continuity without interruptions, even during regional failures. This system ensures that data and user profiles in SageMaker domains remain accessible and secure, protecting the workflow of data engineers and scientists.
The new recovery system is based on two modes of operation: active-passive and active-active. In active-passive mode, the infrastructure remains in the primary region, with near real-time data replication to a secondary region that is only activated when the primary fails. On the other hand, active-active mode allows the system to operate simultaneously in multiple regions, synchronizing data through AWS Step Functions that can be invoked, scheduled, or triggered by events.
To achieve effective implementation, SageMaker uses AWS tools like Amazon EFS for backup, AWS Step Functions to automate recovery processes, and the AWS Cloud Development Kit (CDK) to configure necessary infrastructure. This approach ensures that all instances and user profiles are accurately replicated and restored in case of interruptions.
The improvements in SageMaker promise to increase data security and accessibility, allowing for quick and seamless recovery. This advancement is particularly significant for companies that require continuous availability of their artificial intelligence and machine learning applications, providing a robust disaster recovery solution against natural disasters and technical failures. Investing in business continuity strategies reaffirms Amazon’s commitment to offering a secure and reliable environment for data-driven technological progress.
via: MiMub in Spanish
