IBM Elastic Storage Server (IBM ESS) is a software-defined storage option.
N/A
SANsymphony
Score 9.1 out of 10
N/A
DataCore SANsymphony software-defined storage aims to deliver flexibility, scalability, and cost-efficiency in a HCI platform. Powered by a block-level storage virtualization technology, SANsymphony is designed to provide flexibility to control how data is stored, protected, and accessed. The vendor states users can ensure business continuity with just 2 nodes, easily scaling out to 64 nodes, and achieve productivity for performance-demanding workloads by improving I/O processing and reducing…
Reliable data storage and access is a big challenge in AI applications and IBM ESS is a service you need to solve the problem. If you are building an AI service and you already use the IBM ecosystem for AI compute requirements, ESS can serve as an excellent software based storage solution. Integration with other platforms such as EMR/Databricks is still a challenge and should be improved.
Now that you have virtualized your server's CPU and memory resources, you should look to do the same for your storage. Separation of hardware and software has many added benefits not only for CPUs and memory but for storage as well. Without this solution, we would not have been able to afford black-box type SANs at every location. This allowed us to virtualize over 90% of the server environment saving costs and power consumption/cooling and provides all the features costly black-box SANs, including true-HA (which most SAN vendors don't have). Migration from variant SAN storage and using mixed back-end storage solutions is as easier than ever before because the storage being virtualized.
In the 3 years I have been running this, I have contacted support around 4 or 5 times and that was for minor questions with exception of one time when I was performing an update on the system. And in that one time, they were very timely in assisting me with correcting the problem. Top notch customer support!
IBM ESS is optimized for AI and Big Data usecases while S3 is a general purpose storage solutions. EMR and Databricks have lakehouse/data warehousing solutions for distributed computing but are more optimized for just the big data pipelining solutions and not essentially for AI usecases, especially for inference, when you need to load model artifacts really quickly.
DataCore is far easier to manage as well as deal with when it comes to hardware (as DataCore works with any hardware). It also seems way more affordable.
More uptime - Typical SANs have redundant controllers, redundant power supply's and can make the drives redundant by leveraging RAID-0, 5, 6, and 10. The claim to be HA but they are not. That is because if I spay water all over them or catch the SAN on fire, the storage will go down. With DataCore's solution we have identical systems (maybe even at different datacenters connected with long-haul-fiber) including duplicate storage. So one side of the solution can totally be taken offline by water, fire, etc. and the other side will remain up providing true-HA storage. Because of this, we can upgrade the SANs during the day and still keep storage services running (zero-downtime).
Lower Costs - Ability to use 3rd party hardware which lowers the costs, not only for the initial investment but as storage capacity grows.
All the features one can want - High Availability, Thin Provisioning, Asynchronous and Synchronous mirroring/replication, snap-shotting, continues data protection, deduplication, storage reporting, trending with graphs, centralized console for easier management and many more.
