PowerVC

Chat about PowerVC surprisingly starts with OpenStack, which is the core of PowerVC, and it is thanks to OpenStack that PowerVC has such a wide range of capabilities and evolves so dynamically.

OpenStack

OpenStack is an open-source suite of tools for building and managing cloud computing. It is an IaaS (Infrastructure as a Service) platform that provides access to computational resources, networking, and storage in the cloud computing model.

OpenStack can be used to build both public and private clouds. A private cloud is intended for organizations that want to have full control over their IT infrastructure and manage its resources in a modern way.

OpenStack consists of many modules that provide various cloud services. Each module exposes an API, making it possible to build more personalized solutions.

Each module is responsible for a different functionality of the infrastructure. This division helps isolate competencies in large installations (as shown below, it resembles a traditional IT organizational structure) and helps the OpenStack community manage the product code. The most important modules include:

Compute (Nova) – responsible for managing computational resources such as virtual servers.
Networking (Neutron) – responsible for managing the cloud network.
Block storage (Cinder) – responsible for managing block storage.
Object storage (Swift) – responsible for managing object storage.
Identity (Keystone) – responsible for managing the identity of users and services in the cloud.
Image service (Glance) – responsible for managing virtual machine images.

Thanks to these modules, OpenStack can integrate with many contemporary automation tools.

However, for the IBM Power environment to take advantage of OpenStack’s benefits, deeper integration with PowerVM virtualization technologies was required, something that pure OpenStack did not provide. For this reason, IBM decided to develop PowerVC software, which offers this integration.

IBM Power Virtualization Center (PowerVC) is a solution for managing PowerVM virtualization and private cloud, based on OpenStack. It provides deep integration with IBM Power virtualization technologies and support for IBM AIX, IBM i (formerly AS400), and Linux in OpenStack. It also allows network management within the IBM Power ecosystem and storage space provided by storage arrays.

It can be said that IBM built PowerVC based on OpenStack architecture using OpenStack components, but also offered enhancements and components, the most important of which include:

Managing Physical Power Hosts

Effective management of the IBM Power platform would not be possible without access to physical servers and the PowerVM hypervisor. This is achieved via HMC or Novalink consoles, which PowerVC connects to. These consoles provide information about IBM Power servers, and PowerVC manages Virtual I/O Servers (VIOS), controls the network, and storage. PowerVC allows creating groups (availability zones) from hosts, enabling logical management of virtual machines and applying placement policies for LPARs on physical servers. It decides whether VMs should be clustered on one or spread across a group of servers. Thanks to this functionality, as well as Live Partition Mobility, it is possible to use dynamic management of machine placement based on the current load of the physical server (host).

Network Management in PowerVM

Thanks to integration with PowerVM, PowerVC enables managing network virtualization technologies like SEA (Shared Ethernet Adapter) and SR-IOV (Single-root input/output virtualization). It allows managing virtual networks (VLANs) defined on VIOS and allocating IP addresses to LPARs created on PowerVC. This is possible thanks to RMC technology, which enables dynamic allocation of I/O interfaces to LPARs, and the cloud-init software, which helps configure network interfaces during virtual machine initialization.

Storage and Array Management

Here, PowerVC shines and allows:

Integration with storage arrays from multiple vendors such as IBM FlashSystem and Storwize® family, Pure Storage, EMC PowerMax (VMAX), Enterprise Hitachi Block Storage Driver (HBSD), HPE Storage, IBM System Storage DS8000.
The creation of LUNs, snapshots, flash copies, images, synchronously and asynchronously mirrored LUNs.
Support for Shared Storage Pool (SSP).
Cooperation with the SAN network (Storage Area Network) to create zoning for LPARs using NPIV as an interface for the disk subsystem.
Comprehensive disk management at the PowerVM virtualization and VIOS levels, including detecting disks, removing disks, unmapping disks from VIOSs, and attaching disks to instances/VMs.

Graphical Interface, CLI, and API

PowerVC has a dedicated graphical interface from which most functionalities can be performed. You can manage VM instances, migrate them, expand them with CPU, RAM, or disks, manage arrays, add arrays and their pools, manage LUNs, move LUNs between pools, manage the network and IP addressing, manage users and their roles, and much more. If something is missing in the GUI, you can find it in the CLI terminal. PowerVC provides its own set of powervc-* commands, but there is also a full set of OpenStack commands documented on OpenStack’s website. If you want to go for automation, everything available in the GUI and CLI is also accessible via the API.

The features described above are the most important ones added by PowerVC. Delving deeper, there are even more options provided by PowerVC, but the key point is the ability to work with modern tools on the Power Platform.

More Agile

All the features we get are available from OpenStack. This allows management through the previously mentioned OpenStack APIs, enabling the creation of modern solutions built in line with today’s popular IT methodologies and trends: Agile, Dev-Ops, Git-Ops, CI, and IaaS.

In practice, this means that the Power Platform can be managed with tools like Terraform, which allow automatic creation of virtual servers, network management, disk space configuration, location management, etc. It’s a powerful tool that aids in managing the cloud infrastructure of any provider. We also have access to cloud-init software, which helps configure the server/instance/VM during initialization (deployment) from an image, adapting it to security and environmental requirements. Integration with time series databases and Grafana for resource monitoring is also not a problem. This is just a drop in the ocean of modern “devops” tools we can use on Power platforms thanks to OpenStack API and its community. The principle here is: when we come up with something, there’s a good chance that someone else has already done it and posted it on the internet. What we create can be maintained as IaaS code in GIT repositories, and thanks to continuous integration (CI), we can create pipelines that allow us to deploy changes to resources in a controlled manner, following best practices using tools like GitHub Actions, GitLab CI, or Jenkins.

As you can see, PowerVC with OpenStack brings a lot of freshness to IBM platforms, which have not traditionally been known for dynamic development or integration with agile technologies. Thanks to PowerVC, IBM regains some competitiveness in the modern IT world, and the Power Platform becomes more flexible.

A Spoonful of Tar in a Barrel of Honey

Of course, there’s no rose without thorns, and we must mention the disadvantages of PowerVC and OpenStack. Combining two worlds – legacy and modern – always comes with certain issues. These are also human problems, where we need to get used to something new and change long-established standards. Here are some of these problems:

PowerVC is a product under a non-public license, developing less dynamically than open-source tools supported by the community. When a new feature or change is needed, it must be submitted to IBM developers, hoping it makes it to their sprint.
There are occasional bugs in PowerVC, which are non-critical but annoying, such as communication errors with hosts or performance issues when adding more hosts and HMC consoles.
GUI performance: from a user perspective, it is not a lightweight tool, and sometimes the refresh time for operations is long (e.g., a server is on the list but has been deleted).
PowerVC installation is not the easiest, especially on a customized operating system image.
The cloud approach in PowerVC might also pose problems. You have to compromise in some areas, like server or LUN naming conventions, giving up some old habits, or rebuilding old tools.
OpenStack itself is not lightweight or easy, and the entry threshold is quite high due to the numerous modules.

These are the issues at play when deciding whether or not to use PowerVC. These points can be seen as problems, providing further arguments against its use. Alternatively, they can be viewed as challenges to overcome in order to take advantage of the benefits PowerVC provides.

Conclusions

One of the key arguments for using IBM PowerVC is its comprehensive integration with the IBM Power platform (compute, storage, network, etc.). This platform allows unified management of all these elements, eliminating the need to create custom tools and simplifying daily work. In situations where time is crucial, this becomes a priority.

Another important aspect is the APIs provided by PowerVC and OpenStack. By using these APIs, organizations can create dedicated solutions and standardize their IT infrastructure at a higher level of abstraction, regardless of whether it’s an x86 or PPC platform. Tools like Terraform, cloud-init, Jenkins, and Git enable this.

PowerVC, with its comprehensive functionality and ability to effectively manage Power Platform resources, enables the creation and delivery of a global standard for private cloud infrastructure.

Sources:

PowerVC – Private Cloud on IBM Power Platform