What needs improvement with IBM FlashSystem?

The product could be cheaper.

Customization features must be improved.

The solution's infrastructure technology level could be PCI Express 5 instead of PCI Express 4 for the next version. In addition to SaaS-based enclosures, the solution's expansion options must include NVMe options.

In IBM FlashSystem, data reduction is an area with shortcomings where improvements can be made in the future. In the future releases of IBM FlashSystem, I want to see synchronization in three-site replication specifically. The response time taken by the technical support team of IBM is an area where improvements are required. It takes a lot of time for me to get to IBM's second line of support to get an answer.

IBM could do more marketing and improve brand awareness. I had never heard of this product until a colleague told me about it. FlashSystem could add a few features, but it would probably increase the price. For example, Pure Storage offers instant snapshotting and partitioning. That would be nice to have, but I think the cost would go up.

The solution's pricing is a bit high.

Our model does not support compression or deduplication so it has less memory and the controllers are less powerful. We cannot get maximum capacity out of the storage. Bigger models support these capabilities but it would be nice if entry-level storage did too.

The marketing could be improved. Additionally, awareness in the marketplace could be improved.

The setup could be improved.

There should be more licenses included in the standard price. Additional licenses might be added for the fundamental licenses, such as those for copying and flash copies.

This solution could be improved by offering greater amounts of storage.

The only issue my team faced was transferring the data from the old system to IBM FlashSystem, which is an area for improvement in the solution.

The solution has a low number of NVME host attachments at 16 per IO group over the fiber channel. This is magnitudes lower than competing products. The 8.5 release for the 7300 and 9500 Flash Systems no longer allows IO group migrations. The replacement volume mobility is not as seamless as IO group migrations. The Kubernetes CSI driver and the open-stack cinder driver still rely on SSH instead of native APIs for configuration changes. This reduces the limit of outstanding configuration changes that can be submitted to storage in bulk. The solution has not yet adopted Swordfish APIs and its SMI-S APIs are legacy and depreciated. Swordfish's are vendor-independent APIs made by the Storage Industry Association that allow you to manage storage no matter your vendor. These new generation APIs were released after ten years but IBM has not yet jumped on board. With a multi-vendor environment like ours, implementations are easier with universal APIs. Redhat Enterprise Linux clones such as CentOS, AlmaLinux, or Rocky Linux are not supported. All are binary compatible and should be supported because they are fundamentally the same product with different branding. It would be helpful to have a public page listing the minimum supported firmware levels for HBAs from different vendors. We have run into bugs with fiber channel cards that were solved with firmware updates. It was a laborious process to cross-reference vendor information so it would be helpful for IBM to provide recommended baselines for firmware.

The interface could improve in IBM FlashSystem.

The generic functionality of IBM FlashSystem, IBM always dismisses using file share or sharing protocols inside their storage hardware, and they only focus on the block-level storage.

The deduplication and compression ratio is not very good. It's not reaching a very high ratio. I would like to see the replication or synchronization between the storage to be IP replication. Right now, the IP replication is only supported in this model. So, the pure IP will be okay.

This solution is expensive compared to other systems. The interface of this solution could be improved.

This product lacks some of the options we wanted. For example, expansion was difficult and it required a lot of patching to be done. It's not a seamless process because you need to do multiple modifications or alterations. There is a lot of effort required by the customer in order to expand the hardware. Further expansion of our model is not feasible. As new models are released, expansion becomes difficult and you have to switch to a new model when you want to upgrade. This means that you have to migrate data between models. We want something where you can use for at least five years, where you can expand the hardware without the inconvenience of changing models. Ideally, the operating system and other applications would not be affected when expanding. It would be helpful if the solution had built-in safeguards against security threats and malware, such as ransomware. Anything that can be utilized to enhance data integrity would be helpful.

The pricing could be improved, but I think it's getting better and better with each version. IBM needs to implement NAS storage again, as this is a big flaw. Dell EMC is very good at this and if you compared them based on NAS storage, Dell EMC would win right away. IBM's solution for NAS storage is very complicated. We don't have a storage box that provides file sharing from itself, we have to put software on it and go through a whole complicated process. It should be simplified.

The price is very costly, which could be improved. In the next release, I would like to see some integration with VMware so that storage can be managed directly from a single pane.

The Data Reduction Pools (DRP) support could be better.

Cloud file sharing is an area that needs improvement.

We had issues when attempting to do a flash, we hope to resolve it soon.

Recently, we deployed SS9100. At the core level that is deployed on that storage, it is not stable. We had an incident not too long ago. Both controllers rebooted simultaneously, within 15 seconds. There was some threshold value defined in the core level, and the system exceeded that threshold value. We logged the case to IBM. IBM did internal checks, which we deployed. The permanent fix will be available in the first quarter of 2021. It seems to be an issue on IBM's side. Obviously, we were surprised by how both controllers rebooted. We faced downtime on our applications and on our services. The issue which we had recently faced relates to the core level. It should be first tested at IBM labs and then introduced for general release. IBM should improve its data reduction development.

Their technical support needs improvement in terms of reachability for the clients and response times. They should be more responsive and have more online platforms for support. They should make more technical information available online. There could be some kind of documentation community.

The stability of the solution isn't great. We have had a lot of issues with discs over the years. There should be better integration with utilization platforms. The pricing needs to be more competitive.

They can improve its initial configuration. The initial configuration is currently very difficult. There are multiple choices or alternative ways to configure based on the use case and what you are targeting out of the device, that is, more capacity or more performance. These multiple alternatives cause a lot of confusion. They should increase the processing part of the nodes. Currently, you can cluster up to eight nodes. From my experience and the workload that I am facing in my environment currently, I would like to see either a bigger or stronger node or a larger number of nodes that can be clustered together. We formally communicated to them that we need to see either this or that, and they are working on something.

It could be easier to implement. This is an expensive product and if the price were reduced it would be better.

There is always room for improvement, but IBM is less interested in on-premise storage or on-premises solutions going forward. They're highly focused on the cloud. I don't see IBM being a very major player, the way that they used to be because they are moving away from this and are trying to move all their customers to the cloud. Nothing really comes to mind for needing improvement. Some years ago, there would have been an answer to what could be better about this product, but nowadays, virtually all of the companies are meeting all of the features. More and more, we would like to see how it's easier for the solutions to be bought by the customer more on a pay-per-use basis. That is certainly an improvement. The customer's expectations are what they get on the cloud, they're expecting even in the on-premises deployments, going forward. They want to pay-per-use and not own and get stuck with what they're buying. They want flexibility. IBM does that in a few products, but then more and more you see the business model changing towards that. We'd like to see that in all IBM products.

They can include Amazon file system S3 protocol in the upcoming releases. It is a cloud file system. IBM FlashSystem doesn't have this feature in the box for high-end or mid-range. We have got requests for this from customers because we need to use S3 for EDI application storage. At the beginning of every year, IBM releases firmware. When I find any bugs in the firmware during the year, I am unable to find any information from IBM regarding the bug. I need to open a ticket, and the IBM engineering team makes a patch only for me. This patch is not public. By creating a customized patch for a client, they don't really solve the issue for everyone. If multiple users have the same bug, IBM should upload the patch on the official website so that we can download it. IBM FlashSystem has a monitoring tool in the box, but it is not advanced. I need a more advanced tool for more advanced equations and monitoring. All top three storage vendors, that is, EMC, IBM, and Pure Storage, don't have a powerful monitoring tool. To monitor our box to show the statistics for I/Os and latency, I need to pay for extra software. The built-in monitoring storage is not mature enough to handle all requests and generate all reports that I need. They can include the functionality to stretch a cluster natively without using any additional boxes. In addition, there are some features that EMC has integrated with the box. These features are not available in IBM FlashSystem.

When you add more flash drives or more NVMe, that is the bottleneck or shortcoming with this type of storage. When you fill the first enclosure, you won't be able to add more NVMe until you add more SAS drives. The security features can be improved such that the encryption does not affect performance in any way.

Enterprise data storage needs improvement. They should create a feature for data and file storage.

The ease of installation should be improved. We had issues with the configuration model. In the next release, there should be and flash and caching features. Customers also have problems accessing their files from the storage. That's what they usually complain about. This is something they should improve. The main issue is the speed in terms of accessing the data. That is the customer's big complaint. They also complain about the speed of the hard drive.

I would like to see an improvement in the handling of large amounts of rights. An automatic flash system that doesn't do compression or deduplication will flush through the rights directly from the host to the flash modules. It doesn't keep them in the cache. For compression and deduplication systems, they have to do compression, deduplication and the memory and cache for the controller. So they have to keep the data there otherwise you will find yourself stuck with performance issues.

ISCSI/ISER Bandwidth is at 25gbps. Would be nice if that can keep track with the bandwidth currently available in new dc networking deployments.

The support is simply not there, so it needs to be improved. This solution needs a management console where we are alerted to issues and can report them, or escalate them through email or another method. If something happens to our storage, for example, then we will be notified, and we can report it through the console.

They should reduce the cost of the solution and offer smaller storage sizes to target small to medium-sized businesses. The solution is not easy to implement. It takes a lot of time to study the product and it's a little complicated in general. The solution should offer integration with the cloud on the next upgrade. It might already be there but we have not integrated that, or gone for that upgrade. It would mean we could keep a copy of our data in the secured cloud, whether it's Amazon or Azure.

Deduplication feature is not mature and it is significantly increasing the latency. Additionally, the product has limitations in the software features and granularity. For example, data reduction feature cannot be aligned at the logical unit (LUN) level. It is aligned at the pool level. However, I expect to create some LUNs with deduplication and HW compression while some other LUNs with HW compression only. The other limitation is available at the remote copy feature. You can not create one-to-many scenarious for the LUNs which is created as the Hyperswap LUN.

- CLI though intuitive, no other API available - Lacks scheduling or a cron - Has no built in short term performance graphs - IBM TPC overexceeds the montoring needs, had to fall back on Stor2RRd - Support response times are bad -

The cluster should be improved because non-disruptive failover was supported only on a few operating systems.

IBM gets a lot of input from the field, from system engineers, regarding what needs to be done and I believe they have a roadmap to try to constantly make the user interface more mature, as well as the options and tools. I trust they are doing good job on this. I have looked at a few pages of a report I download and I saw a graph there regarding software-defined vendors. IBM is not in a good position on this graph. I know that they are working very hard on this, to make it much better and to get to a level where it's not only hardware but also software to provide a complete solution.