{"id":3653,"date":"2011-01-14T00:14:06","date_gmt":"2011-01-14T00:14:06","guid":{"rendered":"http:\/\/blog.open-e.com\/?p=3653"},"modified":"2025-07-03T08:34:21","modified_gmt":"2025-07-03T08:34:21","slug":"what-you-can-expect-from-ssd-2","status":"publish","type":"post","link":"https:\/\/www.open-e.com\/blog\/what-you-can-expect-from-ssd-2\/","title":{"rendered":"What you can expect from SSD?"},"content":{"rendered":"

\t\t\t\tHard disk performance is limited by mechanical parts velocity and there is no solution for this issue.<\/p>\n

Servers are using RAID for redundancy and for performance. Thanks to RAID technology it is possible to scale single hard disk performance.<\/p>\n

The team of disks works faster, but applications are needing increased improvements as over periods technical evolutions.<\/p>\n

There are storage offers using DDR RAM or SLC Flash that claim 2-20 GB\/s throughput or 100K to millions of random IO\/sec<\/p>\n

But the cost of such systems are extremely high.<\/p>\n

Recently RAID controller vendors like LSI and Adaptec offer SSD cache options. If the applications require higher random IO\/sec such solutions are worth looking into.<\/p>\n

Adaptec claims 8 times faster IO than HDD-only arrays.<\/p>\n

LSI claims even 50 times faster IO. They have simulated web servers that re-read hot spot activity. With CacheCade enabled reached at 14,896 IO\/sec and compared to 273 IO\/sec with disabled CacheCade.<\/p>\n

Such SATA\/SSD hybrid can significantly\u00a0 improve applications with random IO patterns. They work with single SSD as a minimum , the solution is relatively inexpensive.<\/p>\n

Adaptec\u00a0maxCache\u2122 64GB SSD Cache Performance Kit SRP = $1,795<\/p>\n

And \u00a0LSI CacheCade software pack has a suggested retail price of $270 plus X25-E Extreme 64GB cost $700 makes about $1,000.<\/p>\n

Very interesting will be to know what kind of random IO\/ps can be reached if we use SSD only.\u00a0 We have created RAID 5<\/a> with 4 * INTEL X25-M SATA SSD. The RAID 5 was fully initialized then the IOmeter test was started.<\/p>\n

Results below:<\/p>\n

RAID 5 with 4* INTEL X25-M SATA SSD 120GB<\/p>\n\n\n\n\n\n\n\n\n\n
Transfer Request Size<\/strong><\/td>\nPercent Sequential<\/strong><\/td>\nPercent Random<\/strong><\/td>\nPercent Write<\/strong><\/td>\nPercent Read<\/strong><\/td>\nMB \/ second<\/strong><\/td>\nIO \/ second<\/strong><\/td>\n<\/tr>\n
2KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n17<\/strong><\/td>\n8707<\/strong><\/td>\n<\/tr>\n
2KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n40<\/strong><\/td>\n20266<\/strong><\/td>\n<\/tr>\n
4KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n29<\/strong><\/td>\n7259<\/strong><\/td>\n<\/tr>\n
4KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n56<\/strong><\/td>\n14266<\/strong><\/td>\n<\/tr>\n
64KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n99<\/strong><\/td>\n1580<\/strong><\/td>\n<\/tr>\n
64KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n220<\/strong><\/td>\n3524<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

In order to be in a position to appreciate the above results we have run the same test pattern with RAID 5 with 12 * HDD SAS 15K (SEAGATE ST373455SS )<\/p>\n

Results below:<\/p>\n

RAID 5 with 12* SAS 15K<\/p>\n\n\n\n\n\n\n\n\n\n
Transfer Request Size<\/strong><\/td>\nPercent Sequential<\/strong><\/td>\nPercent Random<\/strong><\/td>\nPercent Write<\/strong><\/td>\nPercent Read<\/strong><\/td>\nMB \/ second<\/strong><\/td>\nIO \/ second<\/strong><\/td>\n<\/tr>\n
2KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n4.4<\/strong><\/td>\n2300<\/strong><\/td>\n<\/tr>\n
2KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n4.8<\/strong><\/td>\n2470<\/strong><\/td>\n<\/tr>\n
4KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n8.4<\/strong><\/td>\n2220<\/strong><\/td>\n<\/tr>\n
4KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n10<\/strong><\/td>\n2630<\/strong><\/td>\n<\/tr>\n
64KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n72<\/strong><\/td>\n1140<\/strong><\/td>\n<\/tr>\n
64KB<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n0<\/strong><\/td>\n100<\/strong><\/td>\n98<\/strong><\/td>\n1570<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

So:<\/p>\n\n\n\n\n\n\n
Random Write IO\/sec with 2k block is<\/td>\n3.8<\/td>\ntimes better<\/td>\n<\/tr>\n
Random Read IO\/sec with 2k block is<\/td>\n8.2<\/td>\ntimes better<\/td>\n<\/tr>\n
Random Write IO\/sec with 4k block is<\/td>\n3.3<\/td>\ntimes better<\/td>\n<\/tr>\n
Random Read IO\/sec with 4k block is<\/td>\n5.4<\/td>\ntimes better<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

and:<\/p>\n\n\n\n\n\n\n
Random Write MB\/sec with 2k block is<\/td>\n3.9<\/td>\ntimes better<\/td>\n<\/tr>\n
Random Read MB\/sec with 2k block is<\/td>\n8.3<\/td>\ntimes better<\/td>\n<\/tr>\n
Random Write MB\/sec with 4k block is<\/td>\n3.5<\/td>\ntimes better<\/td>\n<\/tr>\n
Random Read MB\/sec with 4k block is<\/td>\n5.6<\/td>\ntimes better<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

And the comparison was not quit fair because the HDD array has 12 drives and SSD array only 4.<\/p>\n

We are using cheaper (MLC) non-enterprise version of the SSD. The main problem with MLC SSD is the write endurance which is about 11TB for 120GB drive and 15TB for 160GB.<\/p>\n

If drives are in a RAID 5 the write endurance can be multiplied by n-1. Also there is option to leave some spare space for the Wear Leveling mechanism. For example 10% reserved space expands write endurance by 2.8 times.<\/p>\n

So 4 * 160GB SSD in RAID 5 with 10% reserved space = 3 (n-1)\u00a0 * 2.8 * 15 TB = 126 TB write endurance. If your application is not needing heavy duty write orientations then you can consider it for production.<\/p>\n

If the application is heavy duty write oriented you need to consider Enterprise SSD. Here is the write endurance comparison:<\/p>\n\n\n\n\n\n\n\n
Model<\/strong><\/td>\nEndurance<\/strong><\/td>\nPrice<\/strong><\/td>\n<\/tr>\n
INTEL X25-M SATA SSD 120GB<\/strong><\/td>\n15 TB<\/strong><\/td>\n$400 <\/strong><\/td>\n<\/tr>\n
INTEL X25-E Extreme 32GB<\/strong><\/td>\n1 PB (1024 TB)
\n<\/strong><\/td>\n		$370 <\/strong><\/td>\n<\/tr>\n
INTEL X25-E Extreme 64GB<\/strong><\/td>\n2 PB ( 2048 TB)
\n<\/strong><\/td>\n		$720<\/strong><\/td>\n<\/tr>\n
WD SiliconDrive\u00a0N1x 128 GB<\/strong><\/td>\n701 GB\/day<\/strong><\/td>\n$1500<\/strong><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

How much is 15TB. Sustain write 24\/7 with 20MB\/sec, so 15TB will be written in about 9 days.<\/p>\n

How much is 2PB. Sustain write 24\/7 with 20MB\/sec, so 2PB will be written in about 3.4 years.<\/p>\n

The RAID 5 array with 4 * Intel X25-E 64GB drives will cost $2880 and is about $15 per GB.<\/p>\n

The RAID 5 array with 4 * Intel X25-E 120GB drives will cost $1600and is about\u00a0 $4.4 per GB.<\/p>\n

In comparison to 15k SAS HDD, the RAID 5 array with 4 * Seagate 148GB (15k) makes 4 * $300 = $1200 \u00a0and is about $1.3 per GB.<\/p>\n

This shows about 12 times more expensive but at least 3 to 8 times faster with random IO.<\/p>\n

We did not test the enterprise SSD but mainstream SSD only. The random IO with enterprise SDD would be even better.\t\t<\/p>\n","protected":false},"excerpt":{"rendered":"

Hard disk performance is limited by mechanical parts velocity and there is no solution for this issue. Servers are using RAID for redundancy and for performance. Thanks to RAID technology …<\/p>\n","protected":false},"author":2,"featured_media":55802,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[18,28],"tags":[114,344,388,539,624,625,745,746],"class_list":["post-3653","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-hardware","category-raid","tag-cachecade","tag-iosec","tag-maxcache","tag-raid5","tag-ssd","tag-ssd-cache","tag-x25-e","tag-x25-m"],"acf":[],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/posts\/3653","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/comments?post=3653"}],"version-history":[{"count":1,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/posts\/3653\/revisions"}],"predecessor-version":[{"id":55803,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/posts\/3653\/revisions\/55803"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/media\/55802"}],"wp:attachment":[{"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/media?parent=3653"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/categories?post=3653"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.open-e.com\/blog\/wp-json\/wp\/v2\/tags?post=3653"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}