Everything about Web and Network Monitoring

Home > Server Management > Windows Servers Monitoring > Performance Tuning Windows 2012: Storage Subsystem – Part 3

Performance Tuning Windows 2012: Storage Subsystem – Part 3

Configuration

Performance

Reliability

Availability

Cost, capacity, and power

JBOD

Pros:

·  Concurrent sequential streams to separate disks

 

Cons:

·  Susceptibility to load imbalance

Pros:

·  Data isolation; single loss affects one disk

 

Cons:

·  Data loss after one failure

Pros:

·  Single loss does not prevent access to other disks

Pros:

·  Minimum cost

·  Minimum power

RAID 0

(Striping)

 

Requirements:

·  Two-disk minimum

 

Pros:

·  Balanced load

·  Potential for better response times, throughput, and concurrency

 

Cons:

·  Difficult stripe unit size choice

 Cons:

·  Data loss after one failure

·  Single loss affects the entire array

Cons:

·  Single loss prevents access to entire array

Pros:

·  Minimum cost

·  Minimum power

 

 

 

RAID 1

(Mirroring)

 

Requirements:

·  Two-disk minimum

 

Pros:

·  Two data sources for every read request (up to 100% performance improvement)

 

Cons:

·  Writes must update all mirrors (simplest implementa-tion)

Pros:

·  Single loss and often multiple losses (in large configurations) are survivable

Pros:

·  Single loss and often multiple losses (in large configurations) do not prevent access

Cons:

·  Twice the cost of RAID 0 or JBOD

·  Up to twice the power

RAID 0+1 or 10

(Striped Mirror)

 

Requirements:

·  4 disks minimum

 

Pros:

·  Two data sources for every read request (up to 100% performance improvement)

·  Balanced load

·  Potential for better response times, throughput, and concurrency

 

Cons:

·  Writes must update mirrors

·  Difficult stripe unit size choice

Pros:

·  Single loss and often multiple losses (in large configurations) are survivable

Pros:

·  Single loss and often multiple losses (in large configurations) do not prevent access

 Cons:

·  Twice the cost of RAID 0 or JBOD

·  Up to twice the power

RAID 5

(parity drive)

 

Requirements:

·  One additional disk required.

·  Three-disk minimum

 

Pros:

·  Balanced load

·  Potential for better read response times, throughput, and concurrency

 

Cons:

·  Up to 75% write performance reduction because of Read-Modify-Write

·  Decreased read performance in failure mode

·  All sectors must be read for reconstruction; potential major slowdown

·  Danger of data in invalid state after power loss and recovery if not carefully implemented

Pros:

·  Single loss survivable; active write requests might still become corrupted

 

 

 

Cons:

·  Multiple losses affect entire array

·  After a single loss, array is vulnerable until reconstructed

Pros:

·  Single loss does not prevent access

 

 

 

 

 

 

Cons:

·  Multiple losses prevent access to entire array

·  To speed reconstruction, application access might be slowed or stopped

Pros:

·  Only one more disk to power

 

 

 

 

 

 

 

Cons:

·  Up to four times the power for write requests (excluding the idle power)

 

RAID 6

(Double Parity)

 

Requirements:

·  Two additional disks required

·  Five-disk minimum

 

Pros:

·  Balanced load

·  Potential for better read response times, throughput, and concurrency

 

Cons:

·  Up to 83% write performance reduction because of multiple RMW

·  Decreased read performance in failure mode

·  All sectors must be read for reconstruction; potential for major slowdown

·  Danger of data in invalid state after power loss and recovery if not carefully implemented

Pros:

·  Single loss survivable; active write requests might still be corrupted

 

 

 

Cons:

·  More than two losses affect entire array

·  After two losses, an array is vulnerable until reconstructed

Pros:

·  Single loss does not prevent access

 

 

 

 

 

 

Cons:

·  More than two losses prevent access to entire array

·  To speed reconstruction, application access might be slowed or stopped

Pros:

·  Only two more disks to power

 

 

 

 

 

 

 

Cons:

·  Up to six times the power for write requests (excluding the idle power)

 

 

Certain RAID levels are better than others depending on the purpose of your storage needs. Here are some examples if which RAID level is more suited for which purpose:

·         JBOD configuration: Concurrent video streaming

·         Striping (RAID 0): Data that is temporary or reconstructable

·         Mirroring (RAID 1): Database logs, critical data, and concurrent sequential streams

·         Striped mirroring (RAID 0+1): General purpose combination of performance and reliability for critical data, and high-concurrency workloads

·         Parity disks (RAID 5): Web sites, semicritical data, workloads with large writes, and workloads that are mostly read operations.

·         Double parity disks (RAID 6): Data mining, critical data (assuming hot spares are available), workloads with large writes, and workloads with mostly read operations. RAID 6 is also appropriate for massive datasets, where it would be too expensive to use mirroring and multiple disk failures are a concern.

 

·         Bandwidth and response time improves using more disks.

·         Reliability (mean time to failure) decreases with more disks.

·         Usable storage capacity increases as you add disks, but is more expensive.

·         For striped arrays, more disks mean better load balancing, where fewer disks mean better data isolation.

·         For mirrored arrays, the trade-off is between better cost per capacity (for basic mirrors, which is a depth of two physical disks) and the ability to withstand multiple disk failures (for depths of three or four physical disks). Read and Write performance issues can also affect mirrored array size. For arrays with rotated parity (RAID 5), the trade-off is between better data isolation and mean time between failures (MTBF) for small arrays, versus better cost, capacity, and power for large arrays.

·         Because hard disk failures are not independent, array sizes must be limited when the array is made up of actual physical disks (that is, a bottom-tier array). The exact amount of this limit is very difficult to determine.

 

In our next article we’ll discuss some useful tips for selecting a stripe-size unit, and after that we’ll cover the performance counters. After all, monitoring performance is what we’re all about.

Ard-Jan Barnas

About Ard-Jan Barnas

Ard-Jan is a highly technical writer with deep knowledge into the industry. He has an international background and always brings forth articles that are not just technical but with a mix of business application. This encompassing approach to technology married to business is a welcome approach to writing.

Web & Cloud
Monitoring