Actually, if you look at capabilities of the Promise Pegasus - it's not really that bad. The example I gave includes *2* arrays in a single cabinet, combined worst case total I/O (as measured in Anandtech's review) would come to around 13Gbs....
Another thing that hasn't been mentioned is TB is full duplex; supporting simultaneous Input/Output data transfer (ie: combined 40Gbs).
USB only supports single channel I/O Correction, USB 3.0 is full duplex
Again, choices are good!
Populated with platters it's actually not a bad price at all. And I'll keep my issues with the longevity and support with Promise RAID cards out of this. But where it really reaches it's performance peak which makes it significantly faster than USB3.0 is when populated with SSD's.. and then it gets very pricey. It's a storage solution, so I'll assume the lowest capacity SSD you'd put in there is a 1tb, maybe a 480gb? With current prices you're looking at >$18,000 for SSD's and $999 for the enclosure, for 4tb of RAIDed (10, I find 5 way too risky) storage. So.. $19,000. Will anybody but high end govt and corp supported workstations do this? I doubt many, and at the govt/corp level they wouldn't be doing it with a laptop.. they'd be doing it with a desktop. Build a XL-ATX station with 6 PCIe ports, add 6 Revodrives.. and now you'll have far superior performance. What I'm saying, is by the time we spend for the performance, other options at this price point become more practical.
I'm currently wrestling with the purchase of a Synology 2411+ But not until they equip it with at least USB3, esata, or even thunderbolt. After paying $1800 for 12 3tb platters and $1600 for the case, I'm now at 36tb, or 30tb in RAID 10? And it's not sitting 2 meters away from me shackled by a TB cable. It's tucked back in my office and provides 200mbps (give or take depending on file type) where anyone on my LAN can get at it.. Expensive for an individual user, but actually cheap for a small business with multiple users who needs fast storage. Sure, I could populate it via LAN.. but it now comes with 4 USB2.0 ports you can plug in to. I'd like to see those be four USB3.0 ports, I can imagine it in the middle of a big table at some group workshops being very useful, though I doubt I'd bother. This is very fast storage, much more economical, and any of the current speedy interfaces will work great.
I guess I'm not quite getting the point of using your laptop as a workstation.. where you're plugged into external monitors, storage arrays, etc. Add a small fast desktop stuffed in a Lian-li PC354 case and you've got a more capable solution. Equip it with a Revo3 x2 and 24tb of platters and you have a box marginally larger than this Pegasus storage solution, with more storage, faster performance, but now it's an entire system without the laptop.
All off subject to a degree so I apologize for that, but it illustrates that the PC platform with it's nearly infinite choice of build components continues to provide more economical and ultimately practical solutions. Sure, if I must live with a MBP then a high end storage solution becomes necessary.. but I'm not locked into a MBP.
There's no doubt Thunderbolt is a technically superior interface. But USB3.0 is here now and coming on-line very fast, and it provides all the performance the vast majority of us require. I hope to see TB as an option on future motherboard products and built into storage solutions, but without major intervention from the PC manufacturers I see it going the way of FW800, Mac fiber optic arrays, Expressport, the dodo bird, and every other technically superior interface which wasn't embraced by the PC manufacturers and subsequently the market. Isn't it ironic, that the success of your high end Mac products is so closely (near totally) linked to the success of PC products? Mac needs PC to survive and flourish, but PC would do just fine without Mac..
The last few paragraphs of the AndN review on the Pegasus:
"I also have concerns about cable costs and widespread adoption. For Thunderbolt to really take off we need to see tons of products that support it. Intel's Thunderbolt controller IC can't be cheap, so I am curious to find out if more companies will give Thunderbolt a try. I believe cable costs can be prohibitive, but today device costs are the bigger concern.
Intel already announced that we'd see Thunderbolt support in Ivy Bridge designs next year so it may be at least one more year before we see just how much market potential Thunderbolt has. While I'm happy that Apple is championing the standard, Thunderbolt really needs widespread industry support to make an impact."