I see it, I read it, and I read its rebuttal. Hardly something I'd base facts on, like Chuppa was doing. Again, not quite the citation I was looking for. As it stands, I think Chuppa was wrong.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ThunderBolt Products Starting to Roll Out at NAB
- Thread starter MacRumors
- Start date
- Sort by reaction score
I see it, I read it, and I read its rebuttal. Hardly something I'd base facts on, like Chuppa was doing. Again, not quite the citation I was looking for. As it stands, I think Chuppa was wrong.
It's all rather moot till they get authentic to the standard, and use Fiber Optics!
Substituting old fashioned wire is so misleading.
A bit faster yes, but nothing like Fiber.
Substituting old fashioned wire is so misleading.
A bit faster yes, but nothing like Fiber.
This is rather amusing. I found this Apple Document about Fibre Channel. You will notice that Apple has copper cables available. It would be interesting to compare Fibre Channel with Thunderbolt. Apart from TB integrating video, TB looks a lot like an evolution of Fibre Channel.
As TB progresses the transfer rates should increase which can only be good.
Moot? Whatever.
Hum, you have no idea what Fiber channel is if you seriously claim that. Fiber channel is a networking protocol for storage essentially, Thunderbolt is a host based technology. Call me when Thunderbolt can be switched, redundant, do LUN provisioning and can be extended over a MAN to offer multi-site storage.
Why don't you start reading things before making your incendiary comments?
Reading what ? Someone trying to say TB is an evolution of Fiber Channel when it does nothing of what Fiber Channel is actually used for ? Have you even ever used Fiber channel ?
You really are hopeless! If you can't even figure out a hot link to a source article your opinions are worthless. Go away. Now.
Look, you made a claim that TB seemed like an evolution of Fiber channel. I use Fiber channel day in and out on everyone of my servers. All our storage infrastructure is built off Brocade/HP equipment using a Fiber channel SAN.
If you think my opinion is worthless when I say TB has nothing to do with Fiber Channel, I don't know what to say. Your link does not even begin to explain your comment.
You essentially just claimed USB 3 was a replacement for Gigabit Ethernet. One is a host based interconnect, the other is a networking protocol.
But you're right, I'm probably the one being short-sighted here and not seeing the Apple magic at work.
You're talking about spinning hard drives. Newer SSD drives perform MUCH faster, in fact the fastest right now require direct connection to a high speed PCI-e 8x or 16x. When you start building massive raid and grid arrays, you start reaching a point where you can saturate the line as well.
For a typical consumer this is usually overkill, but for those of us that actually use our workstations for rendering, video editing, heavy data processing, we need this kind of connectivity.
A faster HDD with a faster RPM will add to the speed. With USB 2.0 hard drives, USB 2.0 is the bottleneck. With USB 3.0 hard drives, it'll be the drives. For Thunderbolt, it will be the drives. Would be a while before we see any 10Gbps SSDs (hard drives will never be that fast), so USB 3.0 seems more practical for the time being.
Ummm, look around
Drive arrays can easily exceed 10 Gbps - even with spinning hard drives. Today a 2 SSD drive RAID 0 array can hit 10 Gbps with consumer-grade drives.
Single drives faster than TBolt already exist - 12 Gbps SSD drive 48 Gbps SSD drive.
TBolt devices haven't even hit the market, but TBolt is already too slow for many uses.
Drive arrays can easily exceed 10 Gbps - even with spinning hard drives. Today a 2 SSD drive RAID 0 array can hit 10 Gbps with consumer-grade drives.
Single drives faster than TBolt already exist - 12 Gbps SSD drive 48 Gbps SSD drive.
TBolt devices haven't even hit the market, but TBolt is already too slow for many uses.
Last edited:
Ah, I completely forgot about PCI-Express based SSDs. I kept thinking about 6Gbps SATA SSDs.
Also, "SSD drive" is redundant
.Ah, I completely forgot about PCI-Express based SSDs. I kept thinking about 6Gbps SATA SSDs.
Also, "SSD drive" is redundant
TB is basically PCIe + Display Port video.
Part of the discussion on expansion cards for Mac Pros was that, according to Intel, TB must have direct access to the PCIe lanes and the graphics processor. In theory, someone could come up with a PCIe graphics processor board with a TB controller, ports & etc, but the question will be whether there is a sufficiently large market for the device.
Whether a RAID array saturates TB or not, the bottom line is that TB is a great deal faster than even an eSATA connection. When will TB 2.0 come out? Who knows? Let's see how version 1.0 is adopted.
Cheers
Your arguments would be stronger if you spelled it correctly - "Fibre Channel"
While "great deal faster" is fuzzy language open to interpretation, IMO I don't consider 10 Gbps TBolt to be a "great deal faster" than 6 Gbps SATA. TBolt can't handle two SATA connections at full bandwidth - that's not a "great deal faster" in my opinion.
When it was called "Light Peak", the technology had a lot of promise. Now that it's been downgraded to daisy-chained copper - it's only a little bit better than USB 3.0. Except that we can buy USB 3.0 devices, it's still "in the future" for TBolt devices.
Last edited:
You always seem to forget that Tbolt is a dual bidirectional 10Gb/s channel technology, so in fact it can handle 4 SATA connections, 2 upstream and 2 downstream. With room to spare. On a single port.
But besides the raw speed of TBolt, it's the variety of devices that will be available: high-end storage, audio and video, docking equipment, etc., and the fact that PCIe-class devices will finally be available for computers without PCIe slots. It's perfectly sound for a company like Apple with 90% of their computers without PCIe slots, to pioneer that kind of technology.
Copper or optical wouldn't have change a thing except the max. length of the connection (up to 100m instead of 3m). And FWIW, the only devices you can buy today in USB3 are marginally faster (than FW800) single storage units, and there are already single solid-state drives that are faster than USB3. A couple of video devices that are already obsolete due to some TBolt announcements, and not a single audio interface. No need to trash TBolt because it can't handle $50,000+ devices (RAID Arrays of SSD on 16x PCIe cards) that only a handful of people worldwide will ever buy. As for the "low-end" ioDrive duo, prices range from $9,000 to $12,000 (320/640GB).
Somehow I'm glad Tbolt is not fast enough to handle those devices, I also need a new car.
How does the math work on that? 2 SATA 3.0 connections would saturate the bus right?
check your math
SATA is also bi-directional....
Pointing out that TBolt speed is of the same order of magnitude as SATA speed isn't trashing - it's countering some of the TBolt hype. Some people seem to think that TBolt is infinitely fast.
Agree with you that the real promise of TBolt is to bring PCIe expansion to systems without PCIe slots.
SATA is also bi-directional....
Pointing out that TBolt speed is of the same order of magnitude as SATA speed isn't trashing - it's countering some of the TBolt hype. Some people seem to think that TBolt is infinitely fast.
Agree with you that the real promise of TBolt is to bring PCIe expansion to systems without PCIe slots.
On first glance I was hoping it would be a reasonably cheaper replacement for FC, but that doesn't appear to be the case...