No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>Search

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

SmartTier Feature Guide for Block

OceanStor V5 Series V500R007

This document is applicable to OceanStor 5110 V5, 5300 V5, 5500 V5, 5600 V5, 5800 V5, 6800 V5, 18500 V5, and 18800 V5. This document describes the working principle and application scenarios of the SmartTier (for block) feature. It also explains how to configure and manage the feature.
Rate and give feedback:
Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document. Note: Even the most advanced machine translation cannot match the quality of professional translators. Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
Working Principle

Working Principle

This section provides the concepts and working principle of SmartTier.

Concepts

  • Storage tiers

    In a storage pool, a storage tier is a collection of storage media with the same performance. SmartTier segregates disks into three storage tiers (high-performance tier, performance tier, and capacity tier)) based on their performance levels. Each storage tier contains only one type of disks and adopts one RAID policy.

    Table 1-1 lists the feature of each storage tier.

Table 1-1 Features of storage tiers

Tier

Disk Type and Feature

Applicable To

Data

High-performance tier

SSDs

  • High IOPS
  • Short response time
  • High per-GB cost

Applications with intensive random access requests

Hot data: data that is promoted to a high-performance tier with significantly improved read performance

Performance tier

SAS disks

  • High bandwidth with contending workloads
  • Moderate response time
  • Uncached writes slower than uncached reads

Applications with moderate access requests

Warm data: data that can either be promoted or demoted depending on the precise workload levels and configuration

Capacity tier

NL-SAS disks

  • Low IOPS
  • Long response time
  • High per-I/O cost

Applications with light access requests

Cold data: data that is demoted to the capacity tier without any application performance reduction

  • Storage pool

    A storage pool consists of one or more storage tiers (maximum of three). The types of disks in a storage pool determine the allowed number of storage tiers. However, a storage pool must house more than one type of disks; otherwise, the storage pool can only create a single storage tier and cannot utilize SmartTier to perform intelligent data storage.

    LUNs are created and spread across different storage tiers within the same storage pool before SmartTier is utilized, as shown in Figure 1-1.

Figure 1-1 Logical distribution in a storage pool
  • Data migration granularity

    SmartTier moves LUN data by data migration granularity or data block. Define the data migration granularity when creating a storage pool. Note that this value cannot be changed after being set.

  • Initial capacity allocation

    This function specifies the storage tier to which storage space is allocated during LUN creation.

Working Principle

If a storage pool contains more than one type of disks, SmartTier can fully utilize the performance of each storage tier. With SmartTier enabled, the storage pool identifies the data activity in the unit of data block, and relocates the whole data block to the corresponding storage tier.

TSmartTier manages storage system data at the data block level in three stages: I/O monitoring, data placement analysis, and data relocation. Figure 1-2 displays the flowchart of this process.

Figure 1-2 Three SmartTier stages
  1. I/O monitoring collection is implemented by the I/O monitoring module.

    The storage system considers a data block is "hotter" or "colder" than another data block based on the relative activity levels of the two blocks. The activity level is determined by counting access frequency and I/O ratio for each data block. The storage system uses accumulated weight values to count the activity levels of data blocks. All the data blocks are constantly monitored and analyzed in real time.

  2. Data placement analysis is implemented by the data placement analysis module.

    The module determines the I/O thresholds of data blocks on each tier based on the performance statistics of each data block from the I/O monitoring module, the capacity of each tier, and the access frequency of each data block. The most frequently accessed data blocks (the hottest data blocks) are stored at the tier of the highest performance. Data blocks that exceed the thresholds are ranked. The hottest data blocks are relocated first. The ranking progresses from the hottest data blocks to the coldest data blocks relative to the other data blocks in the same storage pool (note that only the data blocks in the same storage pool are ranked.). SmartTier determines where data blocks are relocated according to the most recent analysis before a relocation.

  3. Data relocation is implemented by the data relocation module.

    SmartTier relocates data blocks according to both the rank determined in the analysis stage and a relocation policy. Data blocks of higher-priority are promoted to higher tiers (usually the high-performance tier or performance tier), and data blocks of lower-priority are demoted to lower tiers (usually the performance tier or capacity tier). Figure 1-3 displays this process..

Figure 1-3 Data relocation

If the space of a higher tier is insufficient and a higher-priority data block needs to be written, an existing lower-priority data block will be demoted to lower tiers to free up the space.

NOTE:

A storage pool configured with SmartTier needs to reserve free space because SmartTier requires extra data exchange space to dynamically relocate data.

When new data is written to a LUN, the storage system allocates the new data to the corresponding storage tier based on the initial capacity allocation policy. If the data activity level changes, SmartTier relocates the data to another storage tier corresponding to the changed activity level and improves storage system performance. The whole data relocation process does not affect the writes of new data. Figure 1-4 shows this change.

Figure 1-4 Data location changes
  1. Initial capacity allocation

    The initial capacity allocation policy determines the tier to which a new data is written. For example, if the Allocate from the performance tier first policy is chosen, new data blocks written into the storage system are placed into the performance tier, as shown in A1 to A6 in the previous figure.

    There are four initial capacity allocation policies available: Automatic allocation, Allocate from the high-performance tier first, Allocate from the performance tier first, and Allocate from the capacity tier first. Automatic allocation means that the storage system distributes new data to the performance tier, the capacity tier, and then the high-performance tier in sequence.

  2. Data relocation

    As the data lifecycle progresses, the activity levels of data blocks A1 to A6 change. The storage system uses accumulated weight values to count and rank the activity levels of all data blocks, and then relocates data blocks according to both the rank and a relocation policy. In the previous figure, data blocks A1 and A2 are of higher activity levels and promoted to the high-performance layer, while data blocks A5 and A6 are of lower activity levels and demoted to the capacity level. The figure shows that A1, A2, and A5 have been relocated to their new tiers, and A6 is in the relocation process.

  3. Writing new data

    The whole data relocation process does not affect the writes of new data. In the previous figure, data blocks A7 to A10 are newly written into the storage system, and they are placed to the performance tier according to the Allocate from the performance tier first policy. During the process, data block A6 has completed its relocation.

  4. Continuous running

    As the activity levels of data blocks A1 to A10 change, new data blocks such as A11 to A14 are allocated to the performance tier. The storage system also keeps intelligently managing the placement of data blocks A1 to A14.

Translation
Download
Updated: 2019-07-11

Document ID: EDOC1000181474

Views: 14455

Downloads: 308

Average rating:
This Document Applies to these Products
Related Documents
Related Version
Share
Previous Next