This chapter describes the IBM Spectrum Virtualize product that is implemented in a cloud environment, which is referred to as IBM Spectrum Virtualize for Public Cloud.

A brief overview of the technology that is behind the product introduces the drivers and business value of the use of IBM Spectrum Virtualize in the context of public cloud services. It also describes how the solution works from a high-level perspective.

This publication describes IBM Spectrum Virtualize for Public Cloud V8.4.3.

This chapter includes the following topics:

Companies are currently undergoing a digital transformation and making architectural decisions that determine how their businesses are going to operate in the next few years. They recognize the value of delivering services by using the cloud, and many use public clouds to some degree.

The role of the cloud is maturing and it is more often being considered as a platform for innovation and business value. The cloud is a key enabler to drive transformation and innovation for IT agility and new capabilities.

Nevertheless, one of the challenges for these organizations is how to integrate those public cloud capabilities with the existing IT infrastructure. Organizations want to retain flexibility without introducing new complexity or requiring significant new capital investment.

Cloud integration can occur between different endpoints (cloud-to-cloud, on-premises to off-premises, or cloud to non-cloud) and at different levels within the cloud stack: infrastructure layer, service layer and for example, at the application layer or management layer. Within the infrastructure as a service (IaaS) domain, storage layer integration is often the most attractive approach for ease of migration and replication of heterogeneous resources and data consistency.

In this sense, coming from the IBM Spectrum Storage family, IBM Spectrum Virtualize for Public Cloud supports clients in their IT architectural transformation and migration toward the cloud service model. It enables hybrid cloud strategies or for a cloud-native workload, provides the benefits of familiar and sophisticated storage functions on public cloud data centers by enhancing the cloud offering.

Running on-premises, IBM Spectrum Virtualize software can virtualize over 400 different storage systems from IBM and other vendors. This wide range of storage support means that the solution can be used with almost any storage in a data center today and is integrated with its counterpart IBM Spectrum Virtualize for Public Cloud, which supports IBM Cloud®, Amazon Web Services (AWS), and now Microsoft Azure.

For more information about Azure, see Chapter 3, “Solution architecture” on page 25, and Chapter 4, “Planning an installation on Microsoft Azure” on page 45.

Designed for Software Defined Storage (SDS) environments, IBM Spectrum Virtualize for Public Cloud represents a solution for public cloud implementations. It includes technology that complements and enhances public cloud offering capabilities.

For example, traditional practices that provide data replication by copying storage at one facility to largely identical storage at another facility are not an option for public cloud. Also, the use of conventional software to replicate data imposes unnecessary loads on application servers.

For more information about use cases, see Chapter 2, “Typical use cases for IBM Spectrum Virtualize for Public Cloud” on page 9.

 

IBM Spectrum Virtualize for Public Cloud delivers a powerful solution for the deployment of IBM Spectrum Virtualize software in public clouds. This capability allows the use of IBM Spectrum Virtualize for Public Cloud on Azure to enable hybrid cloud solutions, which offer the ability to transfer data between on-premises data centers by using any IBM Spectrum Virtualize -based storage system and multiple cloud environments.

With a deployment that is designed for the cloud, IBM Spectrum Virtualize for Public Cloud can be deployed in cloud data centers around the world.

The term virtualization is used widely in IT and applied to many of the associated technologies. Its use in storage products and solutions is no exception. IBM defines storage virtualization as a technology that makes one set of resources resemble another set of resources, preferably with more wanted characteristics.

It is a logical representation of resources that is not constrained by physical limitations and hides part of the complexity. It also adds or integrates new functions with existing services and can be nested or applied to multiple layers of a system.

The aggregation of volumes into storage pools enables you to better manage capacity, performance, and multiple tiers for the workloads. IBM Spectrum Virtualize for Public Cloud provides virtualization only at the disk layer (block-based) of the I/O stack. For this reason, it is referred to as block-level virtualization, or the block aggregation layer. For the sake of clarity, the block-level volumes that are provided by the cloud are exposed as target volumes, and are seen by IBM Spectrum Virtualize as a managed disk (MDisk).

These MDisks are then aggregated into a storage pool, sometimes referred to as a managed disk group (mdiskgrp). IBM Spectrum Virtualize then creates logical volumes (referred to as volumes or VDisks) that are striped across all of the MDisks inside of their assigned pool.

The virtualization terminology is included in the wider concept of SDS, which is an approach to data storage in which the programming that controls storage-related tasks is decoupled from the physical storage hardware. This separation allows SDS solutions to be placed over any storage systems or more generally, installed on any commodity x86 hardware and hypervisor.

Shifting to a higher level in the IT stack allows for a deeper integration and response to application requirements for storage performance and capabilities. SDS solutions offer a full suite of storage services (equivalent to traditional hardware systems) and federation of multiple persistent storage resources: internal disk, other external storage systems, or cloud and object platforms.

In general, SDS technology uses the following concepts:

IBM Spectrum Virtualize for Public Cloud offers a powerful value proposition for enterprise and cloud users who are searching for more flexible and agile ways to deploy block storage on cloud.

By using standard x86 servers, IBM Spectrum Virtualize for Public Cloud can be easily added to cloud infrastructures to deliver more features and functions, which enhance the storage offering that is available on the public cloud catalog. The benefits of deploying IBM Spectrum Virtualize for Public Cloud are two-fold:

IBM Spectrum Virtualize for Public Cloud enhances the public cloud catalog by increasing standard storage, and offering the following capabilities and features that decrease specific limitations:

A volume’s snapshots occur on high-tier storage with no options for a lower-end storage tier. By using IBM Spectrum Virtualize, the administrator has more granular control, which enables a production volume to have a snapshot that is stored on lower-end storage.

Most cloud storage providers have a maximum volume size (typically a few terabytes) that can be provided by a few nodes. At the time of this writing,
IBM Spectrum Virtualize allows for up to 256 TB and up to 20,000 host connections.

Replication features are natively supported, but are typically limited to specific data center pairs and a predefined minimum recovery point objective (RPO). They are accessible only when the primary volume is down.
IBM Spectrum Virtualize provides greater flexibility in storage replication to allow for user-defined RPO and replication between any other system running IBM Spectrum Virtualize.

IBM Spectrum Virtualize for Public Cloud introduces to the public cloud catalog new storage capabilities. These features are available on IBM SAN Volume Controller and IBM Spectrum Virtualize, but are not available by default. These features that are provided on public cloud are related to hybrid cloud scenarios and its support to foster those solutions for improved hybrid architectures:

In a heterogeneous environment, replication consistency is achieved through storage-based replica peer cloud storage with primary storage on-premises. Because of standardization of the storage service model and inability to move its own storage to a cloud data center, the storage-based replica is achievable only by involving an SDS solution on-premises.

In this sense, IBM Spectrum Virtualize for Public Cloud offers data replication between the IBM FlashSystem family, IBM SAN Volume Controller, or VersaStack and Public Cloud and extends replication to all types of supported virtualized storage on-premises.

Working together, IBM Spectrum Virtualize and IBM Spectrum Virtualize for Public Cloud support synchronous and asynchronous mirroring between the cloud and on-premises for more than 400 different storage systems from various vendors. In addition, they support other services, such as IBM FlashCopy® and IBM Easy Tier.

One of the reasons to replicate is to have a copy of the data from which to restart operations if an emergency occurs. IBM Spectrum Virtualize for Public Cloud enables DR for virtual and physical environments, which adds possibilities compared to the software replicators that are in use today that handle virtual infrastructure only.

IBM Spectrum Virtualize enables the possibility to reverse data replication to offload from a cloud provider back to on-premises or to another cloud provider.

IBM Spectrum Virtualize (on-premises and on cloud) provides a data strategy that is independent of the choice of infrastructure, which delivers tightly integrated functions and consistent management across heterogeneous storage and cloud storage. The software layer that is provided by IBM Spectrum Virtualize on-premises or in the cloud can provide a significant business advantage by delivering more services faster and more efficiently, which enables real-time business insights and supports more customer interaction.

Capabilities, such as rapid, flexible provisioning; simplified configuration changes; nondisruptive movement of data among tiers of storage; and a single user interface helps make the storage infrastructure (and the hybrid cloud) simpler, more cost-effective, and easier to manage.

IBM Spectrum Virtualize for Public Cloud helps make cloud storage volumes (block-level) more effective by including functions that are not natively available on the public cloud catalogs and that are traditionally deployed within disk array systems in the on-premises environment. For this reason, IBM Spectrum Virtualize for Public Cloud improves and expands the capabilities of the cloud offering.

Table 1-1 on page 6 lists the IBM Spectrum Virtualize for Public Cloud features and benefits.

Version 8.4.3 of IBM Spectrum Virtualize for Public Cloud is now available on Azure. Block virtualization further uses public cloud infrastructure for various types of workload deployments, whether it is new or traditional.

The following features are supported on the Azure infrastructure:

The Azure infrastructure is an established platform for today’s computing needs. By deploying the IBM Spectrum Virtualize for Public Cloud platform, the features of IBM Spectrum Virtualize further enrich the capabilities of the cloud infrastructure.

Figure 1-1 shows the general layout of IBM Spectrum Virtualize for Public Cloud on Azure.

In Azure, Locally Redundant Storage (LRS) is attached to the IBM Spectrum Virtualize node instances that compose a single node pair (or I/O group) that provides a shared storage pool that is used by IBM Spectrum Virtualize. IBM Spectrum Virtualize supports the following Azure LRS types: