In a DAS configuration, storage array ports are linked directly to servers, whereas in a SAN a switched network infrastructure exists between storage and the servers. The major advantages of using a SAN are increased flexibility and scalability. SANs allow multiple hosts to access disk storage within a single array through a single port, thereby reducing waste and increasing return on investment.
Typically, intelligent disk storage arrays are used in SAN environments. SAN-attached intelligent disk arrays offer additional functionality, including local replication and remote replication, snapshots, data deduplication and storage virtualization, and are highly resilient to failure of a single component.
Conversely, DAS array ports are connected in a 1-to-1 ratio with the server. While intelligent disk arrays can be used, it is more common to see Fibre Channel or SCSI-attached JBOD arrays used in a DAS configuration.
Unintelligent disk arrays are cheaper but do not offer the same levels of performance and scalability as their intelligent counterparts. Expanding DAS capacity can be tricky, requiring additional hardware purchases and outages as extra cards are installed into servers.
If your 10 TB is to be assigned to a single host and if this configuration is not likely to change, then the lowest cost option will be to use a directly-attached unintelligent disk array. However, if your 10 TB is to be distributed between multiple hosts and your data needs to be highly available or has high performance requirements, then a SAN-attached intelligent disk array will result in a much more flexible option.
Related Q&A from David Boyd
Adding drives to a RAID 5 set increases your ROI and the likelihood of simultaneous drive failures, and lengthens RAID rebuild times. So what should ... Continue Reading
For disaster recovery, you can restore backup copies made from data deduplication. However, it is crucial to protect your deduplication indexes and ... Continue Reading