Skip to main content

Dynamic Capacity Allows Businesses to Respond to Changing Demands

IBM Offering Manager George Gaylord explains how flexible compute offerings meet client needs.

Image compilation of people at work in different settings.

IT must be agile to handle shifting business demand. Disruptions can occur for reasons big and small, including seasonal demand, changes in the economy or the acquisition of new customers. On-premise private cloud infrastructure needs to adapt to these changes quickly without busting budgets. Many IBM clients are looking to consumption-based IT to navigate fluctuations in compute and memory requirements.

To remain competitive, businesses need an IT strategy that accommodates the rapidly changing business environment. A consumption-based approach to IT infrastructure helps clients thrive, and it’s made possible by IBM Power Systems™ Dynamic Capacity (ibm.co/2OauNcM). This approach not only helps deliver large-scale, private-cloud infrastructure to mission-critical applications, but also can lower up-front capital expenses. 

IBM’s Dynamic Capacity enables clients to turn on and pay for extra capacity when it’s needed, by the day or even by the minute. An organization can increase capacity for a short time with just a few keystrokes. If a client needs to permanently expand capacity, IBM makes it easy to do so. More importantly, a client may scale back if the additional resources aren’t required on an ongoing basis.  

With Dynamic Capacity, IBM has enabled Elastic Capacity or Shared Utility Capacity on IBM Power® E980 and/or E950 servers across an enterprise, with self-service provisioning of Elastic Capacity and automated resource sharing and access to fully active additional resources, metered by the minute. 

IBM Capacity on Demand Offerings Explained

IBM has offered Capacity on Demand technology for generations on IBM Power Systems enterprise-class servers (ibm.co/2OauNcM), enabling clients to seamlessly activate compute and memory resources as needed on their systems. The offerings are:

  • Capacity Upgrade on Demand: Inactive processor cores and memory units are permanently activated. The client purchases an activation feature, enters the activation code and engages inactive cores and memory. This can be done seamlessly with no disruption to business or need to restart the server.
  • Trial Capacity on Demand: Clients can evaluate their need for turning on inactivated processor cores, memory or both in their Power Systems servers. The trial lasts for 30 days and there is no charge to the client.
  • Elastic Capacity on Demand: Clients can temporarily activate processor cores or memory units on a per-day basis through the Hardware Management Console.  
  • Processor Utility Capacity on Demand: Clients typically choose this option to handle unpredictable, short workload spikes. Utility Capacity on Demand provides additional processor capacity temporarily within a shared processor pool. Usage is measured in processor minute increments.
  • Power Enterprise Pool: This is a group of Power Enterprise Systems that share Mobile Capacity on Demand processor and memory resources. With Power Enterprise Pools 2.0, a client’s servers are aggregated into a pool with all processor and memory resources fully available. Clients pay on a minute-by-minute basis for any resources used above the base level. 

Right-Sized Resources

Traditionally, IT purchased server capacity based on peak demand even though the business only needed a portion of that capacity on a daily basis. For instance, a user requires an additional 10-20% capacity once a month above average daily capacity to run a month-end report. The client might be paying for a lot of unused capacity the other 29 days of the month. “Many enterprises have traditionally had to over-provision server infrastructure to respond to peak processing demands that aren’t required on a sustained basis 24-7-365,” says George Gaylord, program director, Offering Management, Power Enterprise Servers.  

The IBM Power Systems platform now offers a different purchasing model. A client buys a server that has more capacity than needed for current workloads. The client pays only for the activation of cores and/or memory it needs right now. As demand for compute grows, the organization can enable access to additional capacity on the server. For example, a client might purchase a server with 32 cores and 1 TB of physical memory. But the client only activates eight cores and 256 GB of memory at first. The TCO reflects the portion being used by the client initially. 

“It’s always been a balancing act where organizations try to size for peak consumption, but want to optimize and lower their cost of ownership,” Gaylord says. 

“Clients don’t have to bear the full cost of on-premises capacity when it’s only needed in reserve or a portion of time over the life of a project.”
George, Gaylord, program director, Offering Management, Power Enterprise Servers

Finding the Best Fit

As demand grows, the client can choose the best capacity options for the business. The client may want to permanently increase capacity. If so, Capacity Upgrade on Demand would be a good choice. 

Trial Capacity on Demand is used by clients to gain extra capacity for 30 days to meet unexpected spikes, emergencies or to accommodate the needs of a short-term project. This is a one-time opportunity for the client to turn on additional processors to complete a project. 

When spikes in demand occur due to unusual demand or seasonal demand, the client may opt for Elastic Capacity on Demand. This offering is available in most countries worldwide. 

Retail and financial services clients have used Elastic Capacity to meet demand during bigger-than-expected holiday seasons, enabling additional compute capacity only for a couple of weeks, turning it off when the workload decreases after the holiday.

“Clients don’t have to bear the full cost of on-premises capacity when it’s only needed in reserve or a portion of time over the life of a project,” Gaylord notes. “And for clients with multiple servers, Shared Utility Capacity takes this concept to a whole new level. It enables purchased compute capacity to be virtually pooled across multiple servers in a private cloud infrastructure. Idle capacity on one system can automatically offset potential charges for reserve resource being used on another system in that same moment.” 

Startups can benefit from Capacity Upgrade on Demand or Elastic Capacity on Demand. These organizations tend to run lean and don’t have deep pockets. However, they need to ramp up capacity as demand rises. They can do that permanently or temporarily. 

Utility Capacity has a proven track record, having been introduced for processors since 2007. At the outset of the current recession, a global investment management firm saw unprecedented demand from their clients. The firm used Utility Capacity on Demand to manage the wild twists and turns in trading activities. “If they didn’t have this flexible capacity, they couldn’t meet their client demand and would have encountered delays in trading or possible outages,” says Adrian O. Robinson, offering manager, Power Enterprise Systems, IBM.

Fitting Budgets and Fulfilling Needs

A pay-per-use approach enables organizations to prudently spend their budgets while fulfilling capacity needs. This approach means other charges such as software and maintenance can be lower, until capacity is required. A client doesn’t have to license additional Power software until the processors are activated permanently. Likewise, maintenance costs aren’t incurred until the processors are activated. 

In addition, Dynamic Capacity offerings give organizations an extremely cost-effective way to handle disaster recovery. They can purchase all of the production capacity for the main site and deploy minimally active capacity in the secondary site. If a disaster occurs, capacity at the secondary site can be provisioned for the duration of the disaster and then turned off when the primary site is back online.

“IBM’s Flexible Capacity offerings allow clients to minimize total cost of acquisition with lower-priced base systems and capital expenditure budgets,” says Robinson. “It allows them to spread out the cost for infrastructure over time and adjust to IT demand peaks as business changes." 

IBM Systems Webinar Icon

View upcoming and on-demand (IBM Z, IBM i, AIX, Power Systems) webinars.
Register now →