SSD for Dell PE R610?

Question

Solid State Drives

There is:
Dell R610. PERC6i. SAS HDD 15k x 4 pcs.
I decided to pick up and stick 2 more SSD drives. Since SAS SSD is more expensive, I settled on SATA SSD. Asked Dell for options for my server. They answered that in the list of supported (dell-approved) only this one "PN PGPVW SSDR,240G,SATA,2.5,SM843T SSD, 240GB, SATA, 2.5 inch, Height 7mm, Samsung, (SM843T)".
But it's 2014. Not on the market.
The server will be used for vmware esxi, on which there will be a terminal with 1C file and the number of users 5-7.
Suggestions for my request from local dealers are:
SSD Intel S3520 Enterprise Series, 240Gb, SATA, 3D1 MLC, 2.5"
Sequential read 320 MB/s
Sequential write 300 MB/s
Random Read (100% span) 65000 IOPS
Random Write (100% span) 16000 IOPS Endurance
rating (lifetime writes) 599 TBW
MTBF 2,000,000 hrs
Intel 545s series, SATA, 256GB, 3D TLC, 2, 5":
Sequential Read 550 MB/s
Sequential Write 500 MB/s
Random Read (8GB span) 75000 IOPS
Random Write (8GB span) 85000 IOPS Endurance
Rating (lifetime writes) 144 TBW
MTBF 1,600,000 hrs
In terms of TBW, the first is better, in terms of IOPS the second. But besides this, I’m still afraid that the disk itself will not take off if PERC limits it, or rather it will definitely limit it to SATA 3gb, but this will satisfy, if only there were no numbers like 100 MB / s.
Which disc would you choose?
Or maybe someone has the same server and modern SSDs are spinning on it, then which models, tell me, please?

Answer 1

For the indicated tasks, any of these will do, not necessarily an enterprise.
But I would look towards MLC.

it will limit unambiguously to SATA 3gb,

It doesn't affect at all.
You have a virtual machine, TRIM will most likely not work, so it makes sense to make over provisioning of sufficient size in any case.

In terms of TBW, the first is better, in terms of IOPS the second.

Is not a fact.
The operating time here is not a very critical parameter.
IOPS indicate the maximum, and in a particular situation, they can be very different from those declared.
On TLC, good IOPS are often achieved through the SLC cache, and as soon as the cache runs out, they fall below the plinth.

If only there were no numbers like 100 MB / s.

Why don't you like these numbers?