I used to think that DAS (Direct Attached Storage was the obvious winner for MPP data warehousing. But last week Vertica told me that a lot of their customers use SANs (Storage Area Networks). Now I’m not so sure.
There certainly are good arguments on both sides. Generally speaking:
- SANs are pulling ahead of DAS.
- Much of the growth in storage is due to data warehousing.
- MPP (Massively Parallel Processing) is pulling ahead of SMP (Symmetric MultiProcessing) for high-end data warehousing.
- MPP architectures are commonly shared-nothing.
- Shared-nothing entails DAS.
But if you think about it, those facts don’t exactly add up.
I made exactly those points in a post on DBMS2, and some great, detailed discussion has ensued. The rough consensus seems to be:
- DAS is indeed more cost-effective than SANS.
- It’s not as obvious that SANs run into any absolute bottlenecks as we first thought.
- SANs carry a lot of manageability goodness than DAS doesn’t currently have.
What do you think? Are data warehousing efficiencies enough reason to stick with DAS, or are you going with SANs all the way?
