This chapter outlines the benefits that the IBM System Storage N series provides to VMware vSphere 4.1 environments. It includes the following topics:

In a VWware vSphere 4.x environment, the performance and availability of the storage system are important. Generally many different server systems are consolidated onto each VMware ESX host, and a failure can cause all of the machines to have an outage or data loss. RAID-DP (Figure 3-1) provides the benefit of both performance and availability without the requirement to double the physical disks. This benefit is achieved by using two dedicated parity disks. Each disk has separate parity calculations, which allows the loss of any two disks in the Redundant Array of Independent Disks (RAID) set while still providing excellent performance.

Although you can clone guests natively with VMware, cloning from the N series provides significant storage space savings. This type of cloning is helpful when you need to test existing VMware guests. Guests can be cloned at the N series level and use little additional disk capacity due to the deduplication. We explain how this works in 3.9, “Using N series deduplication with VMware”, where VMware guest cloning doubles the required disk allocation.

The N series storage system provides flexibility in the method and protocol used to connect to storage. Each has advantages and disadvantages, depending on the existing solution and the VMware environment requirements.

Traditionally, most VMware scenarios use standard Fibre Channel SAN connectivity. With N series, you can keep using this method if it is already in the environment. However, fiber connectivity can be expensive if new purchases are required. As a result, more environments are now implementing network connectivity methods to storage. Such methods include iSCSI, Network File System (NFS), and Common Internet File System (CIFS), as illustrated in Figure 3-2.

N series storage systems provide a set of features that make the boot from SAN reliable, secure, and cost effective. You can use these features as follows:

Including many hard drives in the aggregate provides improved performance for LUNs created over them. As a best practice, ensure that each LUN is used by a single datastore, thus making them easier to manage.

Similar backup and recovery requirements provide a good criteria when deciding which servers should share the same datastores. Consider having very important servers on their own datastore, so you can take full advantage of N series advanced functionalities, which are implemented on the volume level.

Using Raw Device Mappings (RDM) with VMware ESXi offers the following benefits:

The N series can facilitate these benefits by providing virtual LUNs though flexible volumes (Figure 3-4).

You can easily increase the storage for a Virtual Machine File System (VMFS) datastore by increasing the size of the N series LUN. Then you add an extent on the VMware ESX Server. However, you must complete this process only when all virtual machines stored on the datastore are shut down.

The use of N series functions, such as Snapshot, allow for fast backup of a whole disk volume without using much additional disk space. The backup can then be written to tape or mirrored to auxiliary storage at the same or different location.

Recovery of a disk volume from Snapshot is fast, because the volume is quickly replaced with the Snapshot. If less data is required for restoration, such as a single file or a guest virtual machine disk (files with .vmdk extension), then the restore depends on the backup strategy:

It is important to note that if no other tool is implemented and a volume backup is taken, only the entire volume can be restored. To overcome that limitation, IBM offers the IBM Tivoli® Storage Manager product. This product interacts with VMWare vSphere APIs for Data Protection, formerly known as Virtual Consolidated Backup (VCB) on earlier VMWare versions. When used together, these products can restore on the image, volume, and file levels from a single backup.

For more information, see the following website:

Deduplication is the concept of storing multiple instances of the same information into a single point. Then a pointer is used to refer to it on the next occurrence, so files that potentially might be stored in an environment many times are stored only once. Microsoft Exchange and Symantec Vault are commercial products known for the usage of deduplication.

N series deduplication provides Advanced Single Instance Storage (A-SIS) at the storage level, rather than the application level. Doing this significantly reduces the amount of storage that is used when the same files are stored multiple times. The deduplication process is shown in Figure 3-5.

You can further increase savings by using N series deduplication and compression with the IBM Real-time Compression solution. Compression, which has been around for several years, has not met the strict IT demands for primary storage until now. To solve primary storage capacity optimization, vendors need to ensure data integrity and availability, without impacting performance or forcing IT to change their applications or process.

The IBM Real-time Compression technology meets these requirements with its Random Access Compression Engine (RACE), through an appliance called Real Time Compression Appliance (RTCA). It provides a tremendous reduction in capital and operational costs when it comes to storage management and the additional benefits of less to manage, power, and cool. Additionally, similar to server virtualization, IBM Real-time Compression fits seamlessly into your storage infrastructure. This is done without requiring changes to any processes and offering significant savings throughout the entire data life cycle.

IBM Real-time Compression provides data compression solutions for primary storage, enabling companies to dramatically increase storage efficiencies. IBM Real-time Compression provides the following benefits: