Today’s data centers present a whole new set of operational challenges for IT departments. Virtualization and the sharing of resources have greatly increased management complexity. Technical workload requirements must be carefully evaluated and managed to ensure service levels are met. Workload placements need to be governed by business and operational policies, and Workload mobility only increases the complexity of ensuring that all requirements are met as changes occur. Cloud operating models tackle these challenges by taking the consumerization of capacity to a whole new level.
IT organizations can learn an important lesson about how to effectively manage shared resources by looking to the hospitality industry. Reservation systems are used by hotels (and really any business that relies on shared resources, such as restaurants, airlines and so on) to ensure that assets are optimized in a way that balances customer satisfaction and profitability. Or, in other words, the goal is to strike a balance between supply and demand.
The Importance of Capacity Bookings
Managing capacity in physical and virtual environments is like living in a house or apartment—houses are not commonly shared, and apartment buildings become vacant and filled again with such low frequency that managers can deal with them in a relatively simple way. Clouds more resemble hotels.
In cloud infrastructure, there are few barriers to coming and going, and capacity can be used for whatever amount of time is desired. Furthermore, hosting internal clouds on converged infrastructure, where large blocks of capacity are typically deployed at once, is a lot like managing a very large hotel, with a “revolving door” of new customer demands to deal with. Given the scale and dynamic nature of these environments, properly matching guests to rooms over time is a tremendous challenge, and doing it wrong will anger customers and make poor use of assets, thereby reducing profitability. Add some additional complexity with weddings or a conference (mass onboarding of consumers) and the situation becomes hopeless.
Unfortunately, many IT organizations are ignoring the need for a new management approach. The trend to internal cloud is like someone going from managing an apartment building to managing a hotel, but management systems have not been upgraded and are falling short. Attempting to manage “new school” infrastructure with “old school” tools is like managing a hotel without a reservation system. This approach can result in huge operational challenges such as not having enough capacity on hand to fulfill requests, not knowing where to place new workloads and slow response times to requests when users are demanding increased agility.
Looking to one of the many emerging cloud stacks to deal with this situation may only make it worse, as that breed of solution invariably focuses on enabling immediate requests, not future bookings, and an infrastructure manager typically has no ability to model future demands and do the appropriate forecasting.
Thus, infrastructure teams are left to wildly overprovision capacity in hopes they will have enough to offset any future demand, eroding the potential savings and efficiency associated with operating shared infrastructure in the first place. If they underestimate, the consequences are equally damaging, resulting in performance and SLA compliance issues, or an inability to fulfill the business’s requirements.
The Capacity Reservation Process
The solution to these woes is a capacity reservation system that enables the following:
- Ability to receive and profile workload requests from users, new applications being developed or physical systems being transformed into the virtual environment;
- Ability to assess whether the workload will fit into the environment on the planned deployment date (taking into account trends as well as other confirmed bookings);
- If it fits, the ability to “hold” or “reserve” the capacity so it cannot be consumed by another user/application;
- Ability to forecast capacity requirements on the basis of all reservations and existing workload growth trends to provide accurate insight into future infrastructure needs.
Although this may sound simple, it actually requires a fairly sophisticated reservations management process, as well as very sophisticated predictive analytics to model the future-state scenarios being assessed.
The following diagram provides a detailed flow for the reservation of capacity.
Throughout the reservation process, a booking goes through several key states:
• Draft—demand requirement has been defined and captured.
• Scheduled—a start date (and optional end date) has been associated with the booking.
• Confirmed—the demand has been analyzed against the future-state models of the environment and has been deemed to fit. The future-state models have been updated to incorporate the demand, so it has priority over future booking requests.
• Rejected—the demand does not fit into the target environment.
• Unconfirmed—the demand does not fit, but the policy does not require confirmations, so the demand will be entered into the future-state model. This effectively means that the booking is granted, but that infrastructure managers will need to add capacity to the environment before the start date in order to avoid going offside from a capacity perspective.
• Canceled—The booking passed the technical hurdles, but the action plan to actually make it happen was rejected, meaning it failed to obtain business or ITSM process-level approval.
• Approved—the action plan to realize the booking was approved.
• Committed—the start date of the booking has arrived and the specific actions to realize it have been “locked and loaded” in the appropriate provisioning, orchestration and/or ticketing systems.
• Expired—the action plan was committed but was not executed (for reasons specific to the automated or manual processes being employed), meaning the booking must be either re-created or rescheduled.
• Placed—the action plan was executed, and the new instances are fully operational, signifying the fulfillment of the booking.
Although this process is more involved than the logistics supporting hotel room bookings, it serves the same purpose: ensuring that applications have the capacity they need when they need it, without forcing infrastructure managers to wildly overprovision their environments to deal with uncertainty.
The lynchpin of the entire booking process is the ability to confirm that an anticipated demand can actually fit into the target environment at the desired future point (and all points beyond), and once it is formally booked into that environment, that other workloads cannot “usurp” its capacity before it is actually deployed. This is where the importance of predictive analytics comes into play. It is imperative for organizations to understand that “old school” trending-only technologies are incapable of representing the complex supply and demand models of cloud environments.
The Impact on IT Operations
The impact of accurate capacity forecasting on both suppliers and consumers of capacity is quite significant. It not only allows supply-side infrastructure managers to right-size their infrastructure (saving millions of dollars), but it gives demand-side consumers greater confidence that cloud infrastructure will meet their needs. Proper forward-looking analytics, based on agreed-upon policies, allows infrastructure managers to achieve the agility they need to compete with external cloud providers while delivering a comparable or better level of service. By looking to similar businesses outside IT, the requirements for (and process of) booking capacity becomes clearer.
About the Author
Andrew Hillier is cofounder and CTO of CiRBA (www.cirba.com). CiRBA’s Data Center Intelligence analytics software is designed to automate IT efficiency management at the enterprise level.
Photo courtesy of Hotel Nuevo Vichona