Testing the Limits of the Dell H710 RAID Controller with SSD
-
Very interesting, thanks. That's really good data to have as lots of people often look to push to arrays larger than that. Pure write workloads would be really interesting. I bet that RAID 10 stayed at the same six drive diminishing returns mark. RAID 5, I bet, would show bottlenecking much earlier as that would trigger overhead not seen in reads.
-
Without something like CacheCade can spinning drives even come close without huge numbers of spindles?
-
@Dashrender said:
Without something like CacheCade can spinning drives even come close without huge numbers of spindles?
Nope. That's never been a possibility. A single SSD often outperforms hundreds if spindles. Even 100 HDs in RAID 0 won't catch my single, old, desktop SSD.
-
yeah, ssd > all
-
Considering that, I guess smart tiered storage is really important today?
-
Tiering or caching. Depends on workload. Cache for throughput. Tier for latency.
-
Great blog post, thanks!
-
You are welcome. We are trying to one blog post every month from our lab. If anyone has any scenarios they want tested, let us know. We can grab just about any Dell server in just about any config necessary and use it in our benchmark tests.
-
Would love to see more of these tests done with other controllers! This article is super popular.
-
@scottalanmiller said:
Would love to see more of these tests done with other controllers! This article is super popular.
I might be doing a bit with an old P400 soon.
-
@Dashrender said:
@scottalanmiller said:
Would love to see more of these tests done with other controllers! This article is super popular.
I might be doing a bit with an old P400 soon.
We might have a P410 soon to do it with too. Would be great info to have.
-
Link seems to not work.
-
@creayt said in Testing the Limits of the Dell H710 RAID Controller with SSD:
Link seems to not work.
A bit of necromancy there. . .
