Fusion Drive: Apple jumps on the SSD cache bandwagon
extremetech.com
extremetech.com
Sensational blog post. Extremetech couldn't be troubled to follow through with due diligence, it apparently doesn't have any journalistic integrity, or editorial review. They literally falsely accuse a company (right in the blog title) of releasing an overpriced commonly available product. Then, at the end of the article excuse their horrible research, by starting with the sentance "it’s possible that Apple". Shouldn't Extremetech know what Apple is doing since you just finished writing a scathing blog post about it??
In fact, they are not caching at all. Extremetech completely blew it on this article, either revealing their lack of professionalism, bias, and/or, both.
This was before the unibody models, but I'm sure the process will be the same for the new systems. Oh, and the best part of going to an Apple repair shop was that my warranty remained intact. I never had any issues getting the Apple Store to work on my laptop after the hard drive was replaced.
If Apple prevents people from doing this with the new hardware, then this will be a disappointment. I haven't kept up with the specifics of their new laptops--I know that the RAM is soldered in, but I hope the hard drive will be replaceable.
It's called bundling and price discrimination.
http://arstechnica.com/information-technology/2012/10/apple-...
Should really improve daily usage performance. Makes sense their benchmarks only go to 4GB.
But it's not an SSD 'cache' right? It's two separate disks/units merged by the OS. Not like Intel, OCZ, etc which is an HDD with some SSD chips "before the bus", Right?
http://s2.gizmologia.com/files/2011/06/2011-05-31-ocz.jpg
Or I am missing out?
That's my understanding, and in fact I expect that would make it work better in some ways.
But if that's so - if it's two commodity pieces of hardware managed by some driver-level software - then it would make sense for it to be cheaper than the custom-hardware solutions from other sources. Or at least, we might expect the per-unit cost to Apple to be lower than the custom hardware systems.
So that just makes the alleged overpricing worse.
Your Seagate hybrid isn't the only system of it's kind, it's not the comparison I had in mind.
It's not a new concept, I do it now on my Win7 machine, but the automated part seems handy.
For example, on my Win7 machine I have a SSD as my main drive and a HDD for storage. Since space is at a premium on my SSD I can install applications that I rarely use and files for storage on the HDD. Applications that I want to load quickly I install them on the SSD.
But I have to do all this manually.
Fusion Drive seems to be software in the OS that handles that idea for you. Anything you use constantly will be moved to the SSD for the speed benefit while everything else you rarely use goes to the HDD where speed is not as big a deal.
They're not directly comparable technically, but I'm in the market for a 27" iMac. The cost of the optimum available spec for me just went down and the convenience went up, so I'm a happy camper.
Has this sort of thing really existed on Windows for years or are people just seeing the words "flash" and "cache" and oversimplifying it?
A detailed review + teardown will able to answer these questions.
Writing data to both the cache and the main store is not a fundamental property of caches. Consider, for example, the caches on the microprocessor of an x86 chip. The only time both the cache and main memory will contain the same data is if the trait is set to 'WT', which is very rare.
It's true that (coherent) caches behave in such a way that you can't tell that the data isn't duplicated at each level, but that's achieved by writing back the data when, for example, an external device requests some particular cached data.
If the cache is persistent, and external devices can access the cache as easily as the main store, there's no particular reason to write data back, other than evictions [1], i.e., you should expect the cache to basically be "a software which moves files around"
[1] This isn't strictly true -- depending on the access patterns and the characteristics of the cache and the bus, you might want to opportunistically write data back to avoid having to write back on an eviction. Disclaimer: I've only worked on microprocessor caches and not disk caches, so I'm not intimately familiar with performance traces for disk caches. However, based on simple back-of-the envelope reasoning, it seems to make sense to make disk caches the type of cache you consider not to be a cache at all.
Also, is it really true that data is always read from the SSD? So if a block is requested that only exists on the hard drive, the data must be copied from the hard drive to the SSD, then read from the SSD--rather than simply reading the data from the hard drive?
Intel has SSD caching support built into the Z68 chipset now, it sounds eerily similar, they could simply be using that. Both reads and writes are cached.
http://www.anandtech.com/show/4329/intel-z68-chipset-smart-r...
IE: there's still some special hardware there but the real smarts was in the driver.
I stand corrected. Regardless, this is only a stopgap solution while SSD prices fall. I expect to see it MIA in 2-3 generations.
Apple has consumerized it now.