Dark Reading is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

News

6/12/2008
03:05 PM
George Crump
George Crump
Commentary
50%
50%

Flash Vs. RAM Solid State Disks

As major vendors ready for entry into the solid-state disk (SSD) market with Flash memory systems, don't count out the traditional RAM SSD. Even though RAM SSDs are more expensive per capacity, companies like Texas Memory Systems are seeing continued growth in RAM-based SSD systems. Why? RAM SSDs have two advantages: speed and reliability.

As major vendors ready for entry into the solid-state disk (SSD) market with Flash memory systems, don't count out the traditional RAM SSD. Even though RAM SSDs are more expensive per capacity, companies like Texas Memory Systems are seeing continued growth in RAM-based SSD systems. Why? RAM SSDs have two advantages: speed and reliability.In the SSD market, speed is king and for customers needing to squeeze every ounce of I/O out of their systems, RAM SSD is still the only way to go. The answer is in the numbers. For comparison, a typical mechanical hard disk drive does 4- to 5-millisecond reads and writes and can sustain about 150 to 300 random I/O's per second.

The typical Flash SSD completes reads in about 200 microseconds (0.2 milliseconds) and 100,000 random read I/O's per second; very impressive when compared with disk. In read-heavy applications, you will see a significant performance increase. Writes, however, are as high as 2 milliseconds and can sustain up to 25,000 random write I/O's per second. While you will still see a performance increase on writes with some Flash SSD vs. hard disks, they're most impressive from a read-performance perspective.

RAM SSD, on the other hand, is significantly faster at both read and write operations. It performs 15-microsecond (0.015 milliseconds) reads and writes and 400,000 random I/O's per second. Significant performance improvement can be seen on both types of operations. The challenge with RAM SSD is that you are dealing with smaller capacity -- 128 GB is typical, but smaller sizes aren't uncommon. You are looking for applications that have specific files that can be moved to the SSD; redo logs, undo segments, indices, and frequently accessed tables are great examples.

Flash SSD has another write-related issue; it can only handle so many. The typical range for Flash SSD is around 1 million to 5 million write cycles. For most applications, this is many years worth of writes. Most enterprise Flash SSDs are made up of multiple Flash modules. Having multiple Flash modules is essential to delivering maximum bandwidth and high availability through RAID protection. Flash SSDs aren't a good fit for latency-sensitive, write-intensive applications; for example, accelerating redo logs, undo segments, and enterprise messaging.

For many applications, Flash SSDs will offer significant and affordable performance increases, but when you need more performance or have legitimate concerns about a high-write application, RAM SSDs are the way to go.

George Crump is founder of Storage Switzerland, an analyst firm focused on the virtualization and storage marketplaces. It provides strategic consulting and analysis to storage users, suppliers, and integrators. An industry veteran of more than 25 years, Crump has held engineering and sales positions at various IT industry manufacturers and integrators. Prior to Storage Switzerland, he was CTO at one of the nation's largest integrators.

Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Commentary
Ransomware Is Not the Problem
Adam Shostack, Consultant, Entrepreneur, Technologist, Game Designer,  6/9/2021
Edge-DRsplash-11-edge-ask-the-experts
How Can I Test the Security of My Home-Office Employees' Routers?
John Bock, Senior Research Scientist,  6/7/2021
News
New Ransomware Group Claiming Connection to REvil Gang Surfaces
Jai Vijayan, Contributing Writer,  6/10/2021
Register for Dark Reading Newsletters
White Papers
Video
Cartoon Contest
Write a Caption, Win an Amazon Gift Card! Click Here
Latest Comment: Zero Trust doesn't have to break your budget!
Current Issue
The State of Cybersecurity Incident Response
In this report learn how enterprises are building their incident response teams and processes, how they research potential compromises, how they respond to new breaches, and what tools and processes they use to remediate problems and improve their cyber defenses for the future.
Flash Poll
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2020-25414
PUBLISHED: 2021-06-17
A local file inclusion vulnerability was discovered in the captcha function in Monstra 3.0.4 which allows remote attackers to execute arbitrary PHP code.
CVE-2021-32078
PUBLISHED: 2021-06-17
An Out-of-Bounds Read was discovered in arch/arm/mach-footbridge/personal-pci.c in the Linux kernel through 5.12.11 because of the lack of a check for a value that shouldn't be negative, e.g., access to element -2 of an array, aka CID-298a58e165e4.
CVE-2021-31818
PUBLISHED: 2021-06-17
Affected versions of Octopus Server are prone to an authenticated SQL injection vulnerability in the Events REST API because user supplied data in the API request isn’t parameterised correctly. Exploiting this vulnerability could allow unauthorised access to database tables.
CVE-2021-34825
PUBLISHED: 2021-06-17
Quassel through 0.13.1, when --require-ssl is enabled, launches without SSL or TLS support if a usable X.509 certificate is not found on the local system.
CVE-2021-32944
PUBLISHED: 2021-06-17
A use-after-free issue exists in the DGN file-reading procedure in the Drawings SDK (All versions prior to 2022.4) resulting from the lack of proper validation of user-supplied data. This can result in a memory corruption or arbitrary code execution, allowing attackers to cause a denial-of-service c...