Resolving Data Storage Latency Issues


 

Data storage latency issues happen when there is a time delay in retrieving or storing data to the data storage device. More specifically, latency is the time delay in starting a command to retrieve or store data at the data storage device.

To understand latency issues, you must first understand how much slower a solid state drive (SSD) or hard disc drive (HDD) is compared to the rest of your computer hardware system. Random access memory (RAM) of your computer operates on the order of gigabytes per second which translates to billions of bytes per second. The central processing unit (CPU) of your computer operates with a frequency on the order of billions of instructions per second.

The fastest data storage devices operate on the order of megabytes per second which translates to millions of bytes per second. This means that an SSD or HDD is at often running at least 1,000 times slower than the other computer system components.

This disparity in speeds is a major cause of latency issues. From the time that the CPU issues a command to retrieve data from data storage, as many as 1,000 CPU cycles might pass before the data storage can even begin to execute the command.

This is true whether the data storage is local or remote. In fact, remote storage suffers from additional latency issues introduced as a result of any communications latency. Thus, even large scale efforts to make enterprise data storage more efficient, such as the U.S. government’s Data Center Optimization Initiative, have not solved latency issues even though it has saved nearly $2 billion since 2016.

Latency Issues From Fragmentation

Latency issues can be compounded by fragmented data storage. Fragmentation is the result of files that are broken into fragments when stored. The result is that additional delays are introduced in retrieving those fragmented files from the data storage.

Thus, latency inherent in starting to execute a command to retrieve data is increased by latency in completing a command to retrieve data caused by fragmentation. As a result, commands that should take thousandths of a second to complete end up taking hundredths of a second to complete.

How to Improve Latency

One high latency fix is to defragment the data storage. A fragmented data storage device results in data I/O that comes in short bursts due to the fragments. Once defragmented, the data storage device produces data I/O that is long and continuous. This shortens the time to completely read and write commands, thus reducing the overall latency.

How to Defragment Data Storage

Tools such as Diskeeper® makes Windows computers faster and more reliable by gathering fragments of broken-up files so they can be stored in a continuous segment in the data storage. This provides for faster retrieval since the computer does not need to hunt around the data storage to collect all the fragments during retrieval.

Latency issues that arise from data storage are somewhat inherent. However, they can be minimized by defragmenting the data storage.

Leave a Reply