Storage

08:00 AM
Connect Directly
Twitter
LinkedIn
RSS
E-Mail
50%
50%

Hard Disk: Not Dead Yet

Even with the arrival of flash storage, don't count out the venerable hard disk. Disk still offers a lot of advantages.

Without a doubt, flash is the most important innovation in modern storage. Flash delivers significantly higher performance for randomly accessing data -- the greatest requirement in today’s virtualized datacenters.

But does that mean only flash is viable for modern storage architectures? That’s what flash-only vendors would have you believe. It's time to take a hard look at the pros and cons of flash and disk.

First, let’s consider disk, the more venerable of the two. Hard disk performance is mostly determined by seek times, which, in turn, is determined by rotational speed, measured as rotations per minute or RPM. Access time for a 15K RPM disk is about 2 ms.; it’s double for a 7.2K RPM drive. Keep in mind that while random I/O incurs seek-time overhead, sequential I/O does not, and that hard disk performs well with sequential I/O.

Plus, hard disk delivers loads of capacity for the price: A single 3.5 inch drive stores up to 6 TBs of data, and a gigabyte costs only a few cents.

Perhaps hard disk’s strongest advantage is its age: Because it’s been around for a long time, its reliability and error characteristics (the likelihood of data/sectors being corrupted) are well understood. And, in the decades since it was first introduced, several technologies like parity/checksums and RAID have been developed to mitigate the impact of errors.  

Now, let’s compare disk to flash.

NAND flash is based on solid-state technology. Since flash eliminates mechanical parts, access times are as much as 100 times faster than hard disk. In an environment requiring lots of random I/O, flash leaves hard disk in the dust. However, flash is far more delicate than hard disk. Flash stores data in memory cells, and every time a cell is erased or written to, it degrades. 

The number of program/erase cycles in a flash device is determined by the kind of flash being used -- SLC, eMLC, or MLC. Devices with a higher number of P/E cycles have better endurance. They also come with a higher price tag: SLC has the highest endurance and is also the most expensive version of flash.

The most common way to extend flash’s lifespan is by over-provisioning, over-allocating capacity so the same cell does not get written to more times than its permitted wear level. That makes flash pricey. Already, it’s seven to 20 times more costly than hard disk. With over-provisioning, it only gets more expensive. 

Of course, the semiconductor and storage industries are working at bringing prices down. At least today, savings are coming at the expense of lifespan. As cost-per-GB comes down, the need to over-provision rises. That means that any initial savings erode -- even vanish -- over time.

Because of its relative newness, flash’s error rates aren't as well understood as those of hard disk. Multiple SSD drives have a greater likelihood of failing all at once -- incidents that can mean catastrophic data loss.

As it turns out, flash and disk are perfect complements. The ideal storage architecture leverages the advantages of both.

Compute power is cheap and plentiful. With that in mind, what if random writes could be coalesced into sequential writes leveraging compute and memory? This scenario is exactly what VMware founder Mendel Rosenblum, proposed in this thesis, "The Design and Implementation of a Log-Structured File System."

Sequentializing can improve performance as much as 100 times, delivering up to 40 MByte/s with a single 7.2K RPM drive. Ten hard disks equal the throughput of an SSD device at a fraction of the cost. That makes it possible to get a lot more performance from disk, while leveraging its capacity advantages.

And remember, not all applications need high amounts of flash. Based on a Nimble Storage survey, the working set size (the actual amount of application data that needs to reside on flash) for low sub-millisecond responsiveness is no greater than 10%. And that includes the most performance-intensive applications.

So the next time a storage vendor tells you that disk is dead, remember flash and disk are complementary. Optimal performance and capacity lies with leveraging both.

Radhika Krishnan is the VP of product and alliances at Nimble Storage. Radhika has more than 17 years of experience in the areas of product and solutions management/marketing, as well as strategic alliances and business development. Prior to Nimble, she was responsible for ... View Full Bio
Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Page 1 / 2   >   >>
joreilly925
50%
50%
joreilly925,
User Rank: Apprentice
5/13/2014 | 2:36:41 PM
Re: applications
@StorageRK, That's a useful statistic. It means that the fast flash tier can be quite small. It maybe opens up a 3-tier solution for larger setups with 10% fast flash, 25% slower, cheap MLC or TLC drives, and the rest cheap bulk SATA storage.
StorageRK
50%
50%
StorageRK,
User Rank: Apprentice
5/12/2014 | 1:43:26 PM
Re: applications
Marcia: as the article indicates, applications such as VDI and OLTP that are "performance intensive" have a small working set size ie the %age of application data that needs to be resident in flash to gain sub-ms latencies has been observed to be on average ~10%. 
joreilly925
50%
50%
joreilly925,
User Rank: Apprentice
5/12/2014 | 12:19:13 PM
Re: It's a bit more complicated!
There are now many 256 GB SSDs on the Internet for less than $120 retail. With 3D flash in production, prices should continue to fall.
aditshar1
50%
50%
aditshar1,
User Rank: Apprentice
5/12/2014 | 6:19:06 AM
Re: It's a bit more complicated!
I tried searching some test results and found below table , one notable fact about SSD is once all the blocks of an SSD have been used, then operations get more complicated and drive performance slows.




Source: http://www.computerworld.com/s/article/9230882/
joreilly925
50%
50%
joreilly925,
User Rank: Apprentice
5/11/2014 | 12:09:19 PM
Re: It's a bit more complicated!
Brian, I'd pondered the cloud pricing conundrum a bit too. I figured out it's more like this:

a 2 TB drive retails for $79, so in volume it's around $60. That brings the raw price down to 3 cents/GB. Then compression and deduplication on very large data sets probably give  n effective capacity multiplier of around 10 x on consumer classs data.

That gets us to 0.3 cents/GB. There's the added cost of the drive box, but a 60-drive barebones from an ODM is really cheap. Offsetting this is the replication policy, with 3 replicas trebling the cost.

Finally. all those rad access charges add up to some serious extra cash.


Clearly it isn't quite at 1 cent/GB, but it's a lot closer than you might think.
Brian.Dean
50%
50%
Brian.Dean,
User Rank: Ninja
5/10/2014 | 11:43:18 PM
Re: applications
Marcia, I feel, ever since virtualization made the processor of the data center to be efficient and utilized at a greater extent, the bottleneck has been storage and memory, one example that does come to mind is that of encryption, but if large volumes of data is being encrypted with lower encryption standards, then again storage and memory becomes the bottleneck.
Brian.Dean
50%
50%
Brian.Dean,
User Rank: Ninja
5/10/2014 | 11:18:09 PM
Re: It's a bit more complicated!
Jim, agreed, the storage environment is a complicated ecosystem where lots of different measures needs to be taken into account. Leaving Google drive aside, that's providing 15GB of free (on the surface) storage space to its users, advantages and disadvantages are present, but focusing on the price side, a Cloud storage providers that uses SDS to provide a gigabyte of storage for 1 cent (for example, Yottabyte), when a 1TB consumer HHD costs around $60 -- almost 6 cent per gigabyte, creates savings.

A 6x lower price differences is a big advantage to the consumer/business, SDS ability to provision the allocation of data in an environment where consumer/business tend to overshoot their requirement might be responsible for some of the saving, economies of scale is another, but I wonder whether these two factors would be enough to account for all the saving produced.  
joreilly925
50%
50%
joreilly925,
User Rank: Apprentice
5/10/2014 | 3:28:52 PM
It's a bit more complicated!
Hi Radhika. The performance issue is a bit less favorable to hard drives. Enterprise SSD have performances that are like this, compared to Enterprise hard drives:

Random IOPs 3000x HDD

Sequential MB/S 5x HDD

Cost 3x HDD

Failure rate in 4 years 1 percent SSD versus 16 percent of HDD.

Enterprise SSD have wear life specs good for at least 8 years of heavy use.

 

Consumer SSD are considerably cheaper, but still real fast.

Consumer SSD compared with Enterprise HDD and the 2TB bulk storage drives:

Random IOPS 200x Enterprise and 400x bulk storage drives

Sequential MB/s 1 to 2x Enterprise and 2 to 4x bulk

Cost 80 percent of Enterprise drives but 6x bulk drives

Consumer drives have less wear life, but configuring them as smaller volumes allows the extra space to go into the spare block pool, extending life for years.

My take is that SSD are today the clear replacement for Enterprise drives, and that bulk, sequentially accessed hard drives are a second tier for cold data. This makes Seagate's shingled drives interesting, due to the  (currently 25 percent) lower price.

Consumer drives may have a major role in the Enterprise as an intermediate tier that reduces the amount of expensive SSD required. This is an important issue as Samsung has just announced mass production of 3D flash, which should sharply lower consumer SSD prices.
MarciaNWC
50%
50%
MarciaNWC,
User Rank: Strategist
5/9/2014 | 4:41:16 PM
Re: not dead yet
"That rabbit's got a vicious streak a mile wide! It's a killer!"

The Holy Grail restaurant sounds very cool!

 
Susan Fogarty
50%
50%
Susan Fogarty,
User Rank: Strategist
5/9/2014 | 4:27:24 PM
Re: not dead yet
It's just a flesh wound! Fetch me a shrubbery, one that looks nice, and not too expensive :)

We actually have a restaurant in the next town called the Holy Grail. It's an old church converted into a pub, very cool -- especially considering that I live out in the sticks. They have all these funky drinks on the menu like "Lady of the Lake" and "African Swallow" and they make their own Holy Grail Ale.
Page 1 / 2   >   >>
Cartoon
Hot Topics
6
What's Next For Flash?
Jim O'Reilly, Consultant,  8/18/2014
3
Aspiring To Smarter Storage
Howard Marks, Network Computing Blogger,  8/21/2014
1
Flash Is Overkill For VDI
Michael Letschin, Director of Product Management, Nexenta Systems,  8/19/2014
White Papers
Register for Network Computing Newsletters
Current Issue
Video
Slideshows
Twitter Feed