JoshDW19
Community Hall of Fame
- Joined
- May 16, 2016
- Messages
- 1,077
IT Brand Pulse recently published a TCO case study on archival storage. In the study, they examine three storage arrays and four software-defined storage (SDS) solutions. Based on their metrics, they found two of the SDS solutions had the lowest TCO and the other two SDS solutions were among the highest, yet at the end of the study, they still conclude that SDS is the future of storage. Putting that minor oddity aside, we agree that SDS has advantages over “traditional” storage array vendors, especially when it comes to not having to pay the “branded storage tax” (a term we’re going to borrow, by the way, thanks guys!). Where we differ from their conclusion, however, is threefold:
1) Fundamentally (and perhaps a little philosophically), all storage is software defined. Some storage software just gets purchased in conjunction with hardware and some gets purchased separately. The former gets lumped in with expensive “proprietary hardware” (typically meaning custom ASICs, FPGAs, etc) and the latter with “commodity server hardware” (i.e. – x86 servers). The former is synonymous with costly traditional storage arrays and the latter with more cost-effective “disaggregated” storage on off-the-shelf servers. Of course, the latter judgment doesn’t ever take into consideration the additional costs of less fault tolerant hardware, dealing with hardware/software/firmware incompatibility issues, having separate hardware and software vendors, and all of the additional costs of supporting and maintaining a disaggregated, multi-vendor solution. Unfortunately, IT Brand Pulse’s Case Study doesn’t take those things into consideration either, but it’s hard to fault them since these things can be difficult to quantify without extensive field study or direct experience…
2) There is a third category of storage: Software-Defined Storage delivered on an x86 hardware platform that is purpose-built for storage. This provides the best of both worlds: the peace of mind and simplicity of “traditional” storage, along with the economics of a disaggregated SDS solution. Think of it as traditional storage without the “branded storage tax” (see, I used the term already!), or “disaggregated SDS” with the look and feel of an actual integrated storage product in conjunction with a far superior support experience to boot. This is where true TCO is found.
3) TrueNAS was not included. Yes, our TrueNAS arrays fall into this third category of storage, and we’ve always been vocal that TrueNAS provides the best value (in both up-front and TCO). Now, thanks to IT Brand Pulse’s research, we have yet another platform to substantiate that claim.
Using the same metrics in their case study, we show the five-year TCO of TrueNAS delivers nearly 25% lower five-year TCO than the lowest solution in the IT Brand Pulse study. Also, an equivalent TrueNAS array provides almost 3x lower TCO than EMC Unity 300, which was the highest five-year TCO in the IT Brand Pulse TCO case study (but, we didn’t need a case study to tell us that, did we?).
Here’s the breakdown of the study with TrueNAS included for comparison:
1) Fundamentally (and perhaps a little philosophically), all storage is software defined. Some storage software just gets purchased in conjunction with hardware and some gets purchased separately. The former gets lumped in with expensive “proprietary hardware” (typically meaning custom ASICs, FPGAs, etc) and the latter with “commodity server hardware” (i.e. – x86 servers). The former is synonymous with costly traditional storage arrays and the latter with more cost-effective “disaggregated” storage on off-the-shelf servers. Of course, the latter judgment doesn’t ever take into consideration the additional costs of less fault tolerant hardware, dealing with hardware/software/firmware incompatibility issues, having separate hardware and software vendors, and all of the additional costs of supporting and maintaining a disaggregated, multi-vendor solution. Unfortunately, IT Brand Pulse’s Case Study doesn’t take those things into consideration either, but it’s hard to fault them since these things can be difficult to quantify without extensive field study or direct experience…
2) There is a third category of storage: Software-Defined Storage delivered on an x86 hardware platform that is purpose-built for storage. This provides the best of both worlds: the peace of mind and simplicity of “traditional” storage, along with the economics of a disaggregated SDS solution. Think of it as traditional storage without the “branded storage tax” (see, I used the term already!), or “disaggregated SDS” with the look and feel of an actual integrated storage product in conjunction with a far superior support experience to boot. This is where true TCO is found.
3) TrueNAS was not included. Yes, our TrueNAS arrays fall into this third category of storage, and we’ve always been vocal that TrueNAS provides the best value (in both up-front and TCO). Now, thanks to IT Brand Pulse’s research, we have yet another platform to substantiate that claim.
Using the same metrics in their case study, we show the five-year TCO of TrueNAS delivers nearly 25% lower five-year TCO than the lowest solution in the IT Brand Pulse study. Also, an equivalent TrueNAS array provides almost 3x lower TCO than EMC Unity 300, which was the highest five-year TCO in the IT Brand Pulse TCO case study (but, we didn’t need a case study to tell us that, did we?).
Here’s the breakdown of the study with TrueNAS included for comparison:
They normalized up-front storage costs for all storage solutions surveyed at 250TB and then increased capacity 25% every year for a total of five years, ending up at 600TB and deriving overall TCO from those costs. We encourage you to read their case study, of course, but for those that want the nitty-gritty, here’s the table of their results with TrueNAS included:
Solution | Type | Cost |
IT Brand Pulse TCO Study Findings |