BLOG: Power Systems Support of Hybrid Multicloud Architecture Part 1 – Historical View of Cloud Computing with Power Systems

May 8th, 2020 BLOG: Power Systems Support of Hybrid Multicloud Architecture  Part 1 – Historical View of Cloud Computing with Power Systems

Ron Gordon
Director » Power Systems

The cloud environment has evolved since 2003 and, with its advancement, IBM Power Systems has transitioned to support today’s hybrid multicloud environment. The journey of Power Systems to become an industry-leading, high-performance hybrid multicloud solution is best understood if we look at a little history. This blog is Part 1 of a two-part series and provides a history of cloud infrastructure evolution. It sets the stage for Part 2, which addresses the specifics of how IBM Power Systems supports the current hybrid multicloud market and technology.

Where and When did Cloud Computing Start?

Cloud computing has been around for years. Back in 2004, when you accessed your Gmail account you were using a cloud-hosted application. Then someone came up with the idea to provide computing hardware as a platform service (PaaS) that you customized to your server needs. Then it was expanded by the cloud provider with the addition of the operating system, virtual machines, and other infrastructure software to provide the full infrastructure as a service (IAAS) where all you added was your application. The last capability added to the cloud expansion by offering software as a service (SaaS) for databases and application code. This enabled you to run Oracle, EPIC, Cerner, SAS, and SAP in the cloud to run your business. At the end of this evolution, customers could access a provider’s cloud infrastructure and, for a metered fee, carve out a set of computing resources (processors, memory, storage, network) with a preloaded operating system, virtual machines, applications, and data that could be publicly accessed by employees or customers. You could also choose the level of cloud support you needed by customizing your environment to have a specific level of PaaS, IaaS, or SaaS environment to meet your needs. Customers could use cloud-based applications to support their business needs, including storing data, developing applications, HPC, AI, etc., while providing a high-availability (HA) environment with disaster recovery (DR) capabilities.

Tier 1 cloud vendors emerged with cloud support beginning in 2006 with Amazon Web Services (AWS), followed by Google Cloud Platform in 2008, Microsoft Azure in 2010, and IBM Cloud in 2011. There are also several tier 2 and 3 providers who emerged early on, including Skytap in 2006. At the time, companies were mainly running x86-based systems enabled for Windows and Linux. The idea was that the cloud, theoretically, eliminated the need for organizations to have a data center with managed servers, ever-expanding in-house x86 clusters, and growing in-house storage systems. Cloud transformation would reduce the staffing growth needed to support application development, production environments, HA and DR, etc. Many startups used this approach rather than spending money on implementing and managing a data center of servers and storage on premise, thus avoiding the associated costs. The characteristics of these startups were that they were small in size, had limited capital, and their focus was on development, not running a data center.

Public Cloud Computing for Data Centers

This startup segment and the enterprise segment led to an explosion in the market called cloud computing. Enterprises started moving their development, data, and applications to a public cloud run by service providers including Amazon, Microsoft, Google, and IBM to purportedly save costs versus doing that work on premise in the classic on-premise data center. No management needed, no data center, no servers, no storage, no environmental costs, and easy expandability.

Many embraced this new concept and started to undertake a cloud-only strategy. However, as public cloud usage increased, reality took hold:

  • It was not saving money, in fact, costs were increasing
  • Security and regulations could not always be met satisfactorily
  • SLAs were not being met
  • Corporate data was not as secure as it was on-premise
  • Performance was sometimes an issue
  • Support and issue resolution was not readily available
  • Staffing was still required

This resulted in 37% of the applications being brought back in-house, according to a 2019 IDC report. For many, the cloud-only strategy has yet to be revisited.

IBM Power Systems’ Movement to the Cloud

For Power Systems users of AIX and IBM i operating systems, moving to the cloud normally implied conversion of applications to Linux or Windows since the cloud providers mainly supported x86-based offerings. Migrations to Linux or Windows involved finding new application providers for x86 or re-writing code to a Linux-based environment. This was time consuming, costly, and risk prone. Also, many enterprise and SMB customers saw that converting to x86 based system would mean they would lose the value that Power System provided when running mission-critical applications. Specifically, they would not have reliability, scalability, security, and performance that Power Systems were providing, all of which were crucial to their business.

Let’s skip forward to today. First, IBM Power Systems 922 and IBM Enterprise Class 880 and 950 servers were incorporated into Skytap in 2018 and IBM Power, the IBM Cloud, and GCP in 2019. These systems support AIX, IBM i, and Power Linux environments using the standard interfaces of the IBM Cloud, GCP, and Skytap. This means that IBM Power Systems customers can now have the option of continuing to run their AIX, IBM i, and Power Linux applications on-prem or migrate them and their associated data to a cloud environment on the same servers they were using in-house. For Power Systems users, migration to the cloud did not have to imply re-platforming and re-coding. The experience of the x86 communities were now available to Power Systems users via the public cloud.

Private Cloud Enables Benefits for Mission-Critical Workloads

The public cloud was real and growing, and the revenue to Amazon, Google, IBM, and Microsoft demonstrated that both small and large enterprises were using their public clouds. There was also more news on cloud usage based on cloud experiences. Companies had learned a lot based on their history with a “cloud-only” or “cloud-heavy” strategy. The idea of complementing on-premise, mission-critical workloads with cloud solutions seemed like a better way to provide customer service, data security, performance, and reliability while rendering to the cloud that which could benefit from the original premise of the cloud.

At the same time came the idea that the on-premise systems could be treated like a private cloud, where automated provisioning of processors, storage, networks, and applications was possible using the same tools the public cloud was using. Enter the private cloud. Used in an organization’s own data center, these tools provided the cost savings customers were seeking through higher systems utilization, reduced server counts, lower management costs, self-provisioning, automated expansion, and recovery while maintaining the security and issue resolution that they needed to run mission-critical work.

The public cloud was still seen as effective for smaller workloads, user interfaces to applications, development and testing, staging, HA and DR, and data backup. Some departments or smaller companies could still implement their entire IT needs on the public cloud, which was optimal for that customer market, and that trend continues. But for enterprise customers, they struggled with how to upgrade their environment and provide consistency and seamless integration between their private cloud and the public cloud.

Cloud Strategy

What is your cloud strategy? Public cloud? Private cloud? Multicloud or hybrid cloud? Does your company employ a cloud services-only strategy? Are you considering a cloud implementation for small workloads, large workloads, or backup only? Let me take a pause here so you can review your company’s current view of their cloud strategy and where you are, or are not, with IBM Power Systems’ participation in the cloud. Then, check out Part 2 – Power Systems Support of Hybrid Multicloud Architecture – Public Clouds Become Part of the Hybrid Multicloud World.

For more information on how IBM Power Systems can support your business, please contact your Mainline representative or click here to contact us with any questions..

You may be interested in:

Power Systems Support of Hybrid Multicloud Architecture Part 2 – Public Clouds Become Part of the Hybrid Multicloud World

Security Challenges in a Multicloud Environment

IBM Multicloud Manager Explainer Video

Mainline