Another blow to 1394?

[DrPizza]

<BLOCKQUOTE><font size="-1">quote:</font><HR>"Huge speed deficit??" Come on, it's not going to stay 400Mbps (or even 800Mbps) forever! Silly argument.<HR></BLOCKQUOTE><BR>Sure, it's not gonna stay at the lame-ass speeds it's at forever (though it probably seems that way to anyone waiting for 800 Mbit FireWire...) *but* the specification doesn't allow for anything faster than 3.2 Gbit/sec. SCSI has been roadmapped beyond that, up to 5.12 Gbit/sec (Ultra640), and there's no reason why it can't go further.<P>To go further with FireWire will, er, necessitate dropping FireWire. IEEE 1394 doesn't allow scaling past 3.2 Gbit/sec. So something new will have to be thought up. It might be backwards compatible (maybe) but it won't be IEEE 1394.

 

[resteves]

<BR>Now who is splitting hairs??<P>I don't think SCSI was roadmapped to 5.2 GPS when it was introduced. And what difference does it make if Firewire2 is called IEEE 1594, as long as it works on the same cables etc.<P>And you are missing the point, SCSI may stay around for the very high end user; but most will rather have the advantages of Firewire, (or USB2 if it wins.)<P>

 

[John]

<BLOCKQUOTE><I>Sure, it's not gonna stay at the lame-ass speeds it's at forever (though it probably seems that way to anyone waiting for 800 Mbit FireWire...) *but* the specification doesn't allow for anything faster than 3.2 Gbit/sec. SCSI has been roadmapped beyond that, up to 5.12 Gbit/sec (Ultra640), and there's no reason why it can't go further.</I></BLOCKQUOTE><P>resteves already said it best: was SCSI mapped-out to 5.12Gbps when it was introduced? Was ethernet mapped-out to 1Gbps when it was introduced? Again, silly argument. 1394 has plenty of room to grow.<P>Furthermore, as SCSI gets faster and faster, the cable and connector issues get even worse. SCSI will simply never compete with 1394 or any other "modern" external high-speed bus simply because its design precludes easy configuration, simple hardware, and small cables and connectors. That will never change unless "SCSI-6" turns out to be a high-speed serial interface with dynamic configuration... View image: /infopop/emoticons/icon_smile.gif<P>Oh, and I still wouldn't call 40MB/sec "lame-ass." Everyone reading this, go duplicate a 20MB file on your fastest disk and tell me if it takes less than one second View image: /infopop/emoticons/icon_smile.gif<P>Again, 1394 will not replace SCSI in high-end applications in the near future simply because the benefits of 1394 (and other high-speed serial interfaces) aren't important in those applications. But <I>high-end</I> is the niche market, not the other way around.

 

[Ophidian]

somewhere i read that the newer scsi standerds were designed to be open where a firewire device and chain could be accessed from the scsi bus .... im not sure where i read this though ... i might do some digging and see if i happened to save it somewhere

 

[John]

That sounds interesting...sort of using SCSI as a fat backplane for multiple 1394 chains maybe?

 

[DrPizza]

Where SCSI was mapped out to at the start is kinda irrelevent. There was no upper limit imposed. This isn't the case with FireWire. It has a limit, and not a particularly forward thinking one.

 

[John]

<BLOCKQUOTE><I>Where SCSI was mapped out to at the start is kinda irrelevent. There was no upper limit imposed. This isn't the case with FireWire. It has a limit, and not a particularly forward thinking one.</I></BLOCKQUOTE><P>Show me where IEEE has stated that no 1394-compatible bus will ever go faster than 3.2Gbps. There is never an "upper limit imposed" on the future of any interface, just the <I>current</I> limit of the spec and technoloy.