Manage Learn to apply best practices and optimize your operations.

Common places for data storage bottlenecks in your IT infrastructure

While easing enterprise data storage performance bottlenecks can require a mix of storage tools and IT expertise, identifying the most common areas for storage bottlenecks can be the first step in troubleshooting them.

Enterprise data storage performance bottlenecks that can clog ports, controllers and disk drives require a mix of tools and IT expertise to find and solve. However, identifying the most common areas for storage bottlenecks can be the first step in troubleshooting them.

Let's take look at the most common places for storage bottlenecks and the potential problems they can cause:

1. Storage-area network (SAN) fabric/Front-end ports into the arrays.

Potential problems include:  

  • Oversaturation/overutilization if there's an inadequate number of ports in the array. 
  • Too much oversubscription in a virtual server environment.
  • Improper load balancing across ports. 
  • Contention/traffic overload at interswitch links. 
  • Host bus adapter (HBA) congestion if a single HBA port is overloaded. Virtual servers can amplify the problem. Multipathing -- using multiple connections from a server to storage -- can alleviate it this problem, according to Sean Derrington, director of storage management and high availability at Symantec Corp.

2. Storage controllers. 

According to Marc Staimer, president of Dragon Slayer Consulting, there is a hard limit on how much performance you can get through a standard active-active or active-passive controller.

"Getting to that limit is determined by the number of drives you have, because there's a certain amount of IOPS and throughput per drive. You could have great performance on the controller, but not enough drives," said Staimer. "So, you're not going to get the controller performance; you're going to get the drive performance at the end of the day."

Potential problems include:  

  • Oversaturating the controller with I/O, limiting the IOPS that can be processed from the cache to the array. 
  • Throughput overwhelming the processor. 
  • Overloaded CPU/insufficient processing power. 
  • Inability to keep up with the performance of solid-state drives (SSDs).

Derrington added that the advent of asymmetric logical unit access (ALUA) storage arrays – where one controller is optimized for a LUN while the other controller is non-optimized -- also increases the likelihood of a performance issue. Again, multipathing can be used to balance the I/O across many paths between the host and storage device.

3. Cache. 

Potential problems include:

  • Insufficient cache memory. 
  • Overloading the cache with writes, causing slow performance. 
  • Thrashing the cache by frequently accessing non-sequential data.

According to Akorri Network Inc.'s CTO Rich Corley, "the cache continuously needs to be filled with new data, so you're not getting the benefit of the cache if it's always refilling itself."

4. Disk drives. 

Potential problems include:

  • Too many applications hitting disks.
  • Insufficient number of drives for throughput or IOPS that application requires.
  • Disks too slow to meet performance needs and support a heavy user workload.
  • Disk groups are potential bottlenecks in a "classic" storage architecture, where the RAID configuration runs over at most at 16 disks. A "thin-wide" architecture, which typically has more disks per LUN, and as a result, spreads data over a wider spindle set -- is less prone to become a bottleneck because of the added parallelism, according to Akorri's Corley.

Learn more about Troubleshooting and identifying data storage performance bottlenecks.

Dig Deeper on Data storage management