By Jean S. Bozman
IBM is targeting multicloud storage as a growth area for its IBM Storage systems and IBM software-defined storage products. On Aug. 14, 2018, the company expanded multicloud support across its portfolio of software-defined infrastructure (SDI) solutions.
The announcements build on earlier IBM Storage announcements for products, features and solutions for simplified backup and recovery, object storage through IBM Cloud Object Storage, and support for scalable clusters in cloud environments.
Multi-cloud environments are the result of growing hybrid cloud deployments – spanning enterprise data centers and clouds – and a growing pattern of business units (BUs) tapping multiple cloud services from CSPs.
Companies are leveraging multiple public and private clouds as they migrate IT workloads from data centers, or choose clouds for remote data backup and archiving. The twin ideas of choice and price are driving much of the movement to cloud, hybrid cloud and multi-cloud services.
Business managers and business units influence these “buy” decisions, choosing to consume specific cloud services—and aiming to control business costs associated with IT. However, IT organizations still have responsibility for ensuring that corporate data is protected and managed in a consistent way.
Multicloud deployments need data protection and security to ensure data consistency across an enterprise organization (or government agency). This is especially important for artificial intelligence (AI) and machine learning (ML), when they are applied to analyze extended data-sets across an organization, whether at data is stored on-premises or in the cloud.
With multicloud strategies, customers must cast a wide net when tracking and managing data. Production data must be consistently replicated and stored, and end-user data must be consistent, no matter where that data is located. For efficiency purposes, large data-sets should be stored close to where they were generated, which reduces the need to transfer large data-sets across many networks and clouds.
Customers are choosing multi-cloud deployments for many reasons: governmental compliance for data locality within geographic regions; reduction in network latency when accessing data; ability to tap different clouds to gain cloud-specific capabilities—and a drive to control their operational costs in competitive markets.
However, multi-cloud scenarios bring a new set of challenges for IT organizations, including:
- API compatibility. Each time data-based applications transit from one cloud to another, they must support the APIs from the clouds on which they are hosted. So, API compatibility is a key factor in supporting multi-cloud operations. This is becoming more important as data-sets are often local, near to where they were generated, but end-to-end applications span on-prem and off-prem storage resources.
- Support for backup and recovery across clouds. Awareness of remote storage systems, and knowledge about their data capacity, is needed for consistent and reliable data backup and recovery – locally and remotely. Typically, cloud data gets replicated three or more times to ensure availability and end-user access, ensuring high availability.
- Need for scalability and automation. The growth of data, and the sheer number of data sources that host that data, require new approaches to automation. Automation gives data administrators more efficient ways to tackle repetitive backup and recovery tasks across the enterprise, including the cloud resources that support enterprise data.
The IBM Announcements
Given these market conditions, IBM announced enhancements to its portfolio of products that support, automate and optimize multi-cloud storage. They are intended to address scenarios in which customers data is hosted in multi-cloud deployments.
Here are quick descriptions of the products announced on Aug. 14, 2018:
- Multi-cloud data protection, using IBM Spectrum Protect automated tiering based on data state. Data can be automatically tiered to the cloud for archiving, based on state of the data. IBM Spectrum Protect Plus environment support expansion to include VMware vSphere 6.7 and application support to include IBM DB2 databases, to improve point-in-time recovery (RTO).
- Enhancements to IBM Spectrum Protect malware and ransomware detection that include the hypervisor data, supporting both VMware vSphere and Microsoft Hyper-V.
- Built-in IBM Storage in IBM Z, ZR1 and Linux One Rockhopper 2 systems that supports industry-standard 19-inch racks, reducing data-center space to house the systems and storage. These configurations support high security and high throughput for data-intensive analytics and HPC applications.
- IBM DS 8880 enhanced support for automated data encryption and transparent cloud tiering. Automation of data encryption for IBM’s all-flash DS-series storage arrays reduces admin time for setup and operations.
- IBM Cloud Object Storage (COS) support for a range of capacities, starting at 72 TB for a 1u rackable system, and scaling up to 1 Exabyte for large, scalable systems used for analytics and HPC. Cloud object storage supports multiple data types – leveraging cloud-based storage for scalability and efficiency. IBM will continue to certify other vendors’ storage systems for IBM COS, extending the total available market for COS use by customers.
- IBM Elastic Storage (ESS) support for IBM Spectrum Scale on AWS, which works with IBM cloud object storage for scalable workloads. Now supported on Power8 systems for “bursting” workloads to the AWS cloud; to be extended to Power9 systems later this year.
- Support for IBM’s FlashSystem 9100, announced this summer, will now be included in IBM Storage Solutions blueprints for reference architectures.
IBM is addressing many of these multicloud challenges in the Aug. 14, 2018 storage announcements. However, customers will want to carefully study, and compare, the options they now have for multicloud data-based solutions. Inherent in the multi-cloud value proposition is the element of choice, to move data to different cloud vendors, as data-storage requirements change.
Clearly, IBM will compete with vendors and CSPs for its share of the multi-cloud data storage software business. For example, Google Cloud Platform and Microsoft Azure plan to capitalize on the capabilities of their internal hyperscale networks for automated storage support. Mindful of that, IBM is counting on the following factors, among others, for differentiation in the emerging market for multicloud storage:
- Ease of use, GUI-based management for products that used to require command-line interface consoles
- Automated feature-sets that span multiple clouds to reduce admin time
- Choice of cloud “targets” for backup and archiving across clouds, reducing total storage costs when pay-as-you-go cloud subscriptions are used.
- Automated or embedded encryption of data on IBM Storage arrays prior to transferring enterprise data to off-premises clouds for later retrieval, backup/recovery or archiving.
- Co-selling with IBM Cloud and IBM software-defined storage support for multi-cloud connections to AWS, Microsoft Azure and other CSPs.
We are in the early stages of multicloud storage. Customers are still getting the feel for the trade-offs between on-site control and cloud pricing that are associated with multi-cloud storage scenarios. IBM’s portfolio is broad and deep, in terms of feature-set. Now, it must make its case for multicloud storage products in a way that is clear and concise to its current – and prospective – customers worldwide.