News
11/7/2008
08:52 AM
George Crump
George Crump
Commentary
50%
50%

SSD's Latency Impact

In our last entry we talked about latency and what it was. We also discussed how storage system manufacturers are trying to overcome latency and performance issues of mechanical drives by using techniques like making the drives faster by using higher RPM drives, array groups with a high drive count, short-stroking those drives, wide striping those drives, and increasing the number of application servers

In our last entry we talked about latency and what it was. We also discussed how storage system manufacturers are trying to overcome latency and performance issues of mechanical drives by using techniques like making the drives faster by using higher RPM drives, array groups with a high drive count, short-stroking those drives, wide striping those drives, and increasing the number of application servers for improved parallelism.All of these techniques cost money, are not very green, and in many cases are more expensive than simply using SSD. Not to mention that they don't typically come close to SSD performance. The result has been the existence of standalone, purpose-built SSD solutions like those from Texas Memory Systems, Solid Data Systems, and Violin Memory, or the manufacturer adding SSD in a "drive-like" manner to its current storage systems.

The speed of SSD technology, especially DRAM, changes the latency focus away from the actual storage medium, as it has now been optimized, and onto the storage system's infrastructure, which is suddenly a lot slower than the storage media. For vendors that incorporate SSD into existing drive enclosures, the performance of the shelf itself becomes a problem, the performance of the processors in the controllers becomes a problem, and an incorrectly sized cache (too big or too small) becomes a problem.

Another factor is that the software load on the controller becomes an issue. For the past several years, storage manufacturers have been piling on features to the storage controller like snapshots, replication, data deduplication, and others. All of these features take computing resources away from responding to storage I/O requests, which worsens system latency.

The result is that while the SSD technology going into the solution may be fast, simply adding SSD to your storage system may not dramatically improve performance like it should. Standalone, purpose-built SSD systems offer lower latency because these vendors have built systems from the chip up that are designed to deliver on the low latency of SSD.

In our next entry, we will examine those differences and how storage manufacturers will need to alter their delivery of SSD technology. Then we will wrap up with capacity management on SSDs. At the cost of SSD technology, the only good SSD is a FULL SSD.

Join us for our upcoming Webcast SSD: Flash vs. DRAM...and the winner is?

Track us on Twitter: http://twitter.com/storageswiss.

Subscribe to our RSS feed.

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
Register for Dark Reading Newsletters
White Papers
Cartoon
Current Issue
Flash Poll
10 Recommendations for Outsourcing Security
10 Recommendations for Outsourcing Security
Enterprises today have a wide range of third-party options to help improve their defenses, including MSSPs, auditing and penetration testing, and DDoS protection. But are there situations in which a service provider might actually increase risk?
Video
Slideshows
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2014-8921
Published: 2015-03-01
The IBM Notes Traveler Companion application 1.0 and 1.1 before 201411010515 for Window Phone, as distributed in IBM Notes Traveler 9.0.1, does not properly restrict the number of executions of the automatic configuration option, which makes it easier for remote attackers to capture credentials by c...

CVE-2014-9676
Published: 2015-02-27
The seg_write_packet function in libavformat/segment.c in ffmpeg 2.1.4 and earlier does not free the correct memory location, which allows remote attackers to cause a denial of service ("invalid memory handler") and possibly execute arbitrary code via a crafted video that triggers a use after free.

CVE-2014-9682
Published: 2015-02-27
The dns-sync module before 0.1.1 for node.js allows context-dependent attackers to execute arbitrary commands via shell metacharacters in the first argument to the resolve API function.

CVE-2015-0655
Published: 2015-02-27
Cross-site scripting (XSS) vulnerability in Unified Web Interaction Manager in Cisco Unified Web and E-Mail Interaction Manager allows remote attackers to inject arbitrary web script or HTML via vectors related to a POST request, aka Bug ID CSCus74184.

CVE-2015-0884
Published: 2015-02-27
Unquoted Windows search path vulnerability in Toshiba Bluetooth Stack for Windows before 9.10.32(T) and Service Station before 2.2.14 allows local users to gain privileges via a Trojan horse application with a name composed of an initial substring of a path that contains a space character.

Dark Reading Radio
Archived Dark Reading Radio
How can security professionals better engage with their peers, both in person and online? In this Dark Reading Radio show, we will talk to leaders at some of the security industry’s professional organizations about how security pros can get more involved – with their colleagues in the same industry, with their peers in other industries, and with the IT security community as a whole.