Is Western Digital updating the raptor series soon?

[The_Wolf_Who_Cried_Boy]

Given it's been more than 18 months since the 740GD's introduction it's about due for an upgrade.

Has anyone heard rumors?

 

[RussSchultz]

I'd like to see a 10K/16MB cache version.

 

[Guden Oden]

Very quiet on the raptor front from what I've seen around the web. WD hasn't made any announcements, and knowing the harddrive industry, an announcement can easily preceede actual availability between anything like three to six months or more.

 

[Rur0ni]

I was concerned about this as well, wanting to hold off on raptors until something better came along, raptors have been out for quite some time now.

 

[ANova]

Raptors are overrated for their price. If you're a speed freak run a raid 10 array.

 

[pcchen]

RAID is not going to make your desktop applications any faster. In most cases, you need better access latency, and only higher RPM (or smaller disk diameter) can provide you that.

 

[arjan de lumens]

RAID is not going to make your desktop applications any faster. In most cases, you need better access latency, and only higher RPM (or smaller disk diameter) can provide you that.

It would seem to me that in a RAID1-type array, if you need a single read access to go faster, it would be possible to issue the access speculatively to both hard drives at the same time and just go with the data that arrive first. If the hard disks then run, say 180 degrees out of sync with each other, then the worst-case access time will be reduced from (1 seek + 1 rotation) to (1 seek + 1/2 rotation), effectively giving you a latency benefit similar to doubling the RPM. The effect will be smaller if the disks run less than 180 degrees ot of sync, but it will be there (this applies to RAID1 only; there is no similar benefit to be had from RAID0 or RAID5)

If you need multiple accesses, any decent RAID controller should be able to load-balance accesses across the different harddisks for increased throughput.

 

[pcchen]

It would seem to me that in a RAID1-type array, if you need a single read access to go faster, it would be possible to issue the access speculatively to both hard drives at the same time and just go with the data that arrive first. If the hard disks then run, say 180 degrees out of sync with each other, then the worst-case access time will be reduced from (1 seek + 1 rotation) to (1 seek + 1/2 rotation), effectively giving you a latency benefit similar to doubling the RPM. The effect will be smaller if the disks run less than 180 degrees ot of sync, but it will be there (this applies to RAID1 only; there is no similar benefit to be had from RAID0 or RAID5)

It's very hard to make harddisks to run in synchronized fashion, so to keep the 180 degrees out of sync constantly would not be easy. But yes, RAID 1 has the potential to give slightly better read latency, or better read throughput.

RAID is good for server enviornment not only because the better reliability, but also the potential of better performance in concurrent access. However, in desktop environment it's not as good. But I think it's may be a good idea to use RAID 0 in desktop environment in near future. Because many people do not bother to backup their data (opposed to server environment), so from this perspective, desktop users may actually need better protection.

Another method is to make a logged, incremental "backup" on another harddisk. Because many data loss in desktop environment is not actually due to failed harddisk, but to user error, a logged, incremental "backup" can provide an opportunity for the user to recover from the error. (or maybe virus attack, etc.)

 

[arjan de lumens]

RAID is good for server enviornment not only because the better reliability, but also the potential of better performance in concurrent access. However, in desktop environment it's not as good. But I think it's may be a good idea to use RAID 0 in desktop environment in near future. Because many people do not bother to backup their data (opposed to server environment), so from this perspective, desktop users may actually need better protection.

I don't understand this argument - surely, RAID 0 would give *less* protection by adding another point of failure - if one disk in a RAID 0 array dies, you lose ALL the data on all the data on all the disks. Are you sure you mean RAID 0 and not RAID 1?

 

[pcchen]

Sorry, my mistake. I meant to say "RAID 1"