Love this term. Can't wait to explore this term at facebook.com/Ibmsux
IBM's Evaluator Group has knocked up sales briefing battlecards that aim to kill off competition from Dell EMC, HPE, NetApp and Pure for its new FlashSystem 9100 flash arrays. We've seen them and summarised the main points to be used by Big Blue and its channel sellers when coming up against these foes. We've also asked the …
"It suggested customers should be asked if they really do need to mix block and file access, and affect system performance."
So what if they answered Yes?
I bet if anyone started asking IBM questions to dig deeper and further clarify their competitive attacks, like "how", "why", and "what", they'll flounder.
It affects system performance? Not sure how they justify that other than FUD that most customers don't really like from vendors. And what if the answer is "no, we are looking at block only so there is no such worry." Even if true with some vendors, at least it's an option to run NAS.
Really weak story against some of the competition. Pure has a roadmap for NVMe over Fabric and NetApp has two platforms that already run NVMe over FC and IB already. DMC has a rebooted VMAX with NVMe with a taped on NAS that ticks more boxes than IBM's. IBM has no cloud angle either.
Maybe they should just hurry up and sell this division to Lenovo too.
The V9000 also uses the SVC/Spectrum Virtualize/V7k/V5k code base (though maybe this FS9100 is the successor to the V9k? Looks like it...) The A9000/A9000R run on XIV/Spectrum Accelerate code, while the FS900 (which I believe is the internal fast-flash-storage building block for the V9k, FS9100, and A9k) runs an IBM-i-fied version of the original TMS code base.
I guess you could say that IBM has a current all-flash option no matter which storage array from them you've used in the past, but it's a fair criticism to point out that three entirely different flash storage lines is a little crazy, as each has advantages the other does not.
Really the V9k (and I think the FS9100) are a marketing bundle of the SVC and FS900. If you already own the SVC/Spectrum Virtualize, you can just put an FS900 behind it, turn on EasyTier, and call it a day; unless there's some sort of pricing games that make the FS9100 a cheaper way of doing that, and you have a good way to migrate between the two.
These Storage (data management) vendors show off tech & performance specs that can only be achieved in lab environments, but not in the real world. When will the storage industry be held responsible - say like VW - for lying to customers.
Customers will never achieve the latencies and IOPS advertised, because real world workloads behave different than load generators.
Storage management tools are almost never worth the investment. The further you push a storage system to the red line (performance) the more unpredictable the system will become. That's a fact for every vendor. At the end you need to put in place a NASA style mission control center filled with staff just to keep driving the system near the red line. Email Alerts and automatic 'advisors' will not help you either. You will need a real human to act upon the email telling you that the max.latency threshold has been exceeded.
Therefore customers will buy systems that are so oversized that they will not hit the wall for the next two years. Once you start over specifying the system size to meet vaguely "projected" performance requirements - you won't be able to compare apples against apples anymore (because different storage systems scale differently).
Battle cards are important so the geeks can have a play at the kids table, while over at the adults table the sales guy and the decision makers do the deals. Buying storage has more to do with emotions and trust than sub-ms latencies...
Biting the hand that feeds IT © 1998–2019