15 Aug 4 Reasons You Need Purpose Built Cloud Reporting to Manage Your Cloud
The delivery of IT services is undergoing a massive shift to on-demand clouds; and with it, cloud reporting and management tools must change as well. Look out for the following considerations when managing your clouds.
Consumers today are spoiled. Uber, Instacart, Grubhub – all racing to deliver instant gratification. Businesses have to move faster than ever as companies race to keep up and compete. Thanks to Amazon, developers now expect near-instant provisioning of IT services. That speed is impossible to deliver from a traditional virtual environment controlled by central IT. This top down approach minimizes responsiveness and flexibility in favor of maximizing reliability for a single specific virtual instance.
Why are traditional virtualization management tools a poor fit for managing cloud environments? Simple – those tools were designed specifically for problems associated with managing a highly controlled virtual environment:
- Request resources through a ticketing process so that IT can prioritize all requests;
- Monitor and control operating system configuration drift to reduce performance impacts to each instance;
- Once approved, manually provision and assign virtual machine;
- Centrally managed bureaucratic infrastructure allocation process;
- Contain user sprawl by limiting the amount of resources that are available to the user.
Assuming that you have settled on building a private cloud (likely on OpenStack), here are 4 reasons why you need a solution designed to solve the management scenarios you will face.
- Self-service is the new consumption model – In a traditional virtual environment, IT resources were hard to get. The process mimicked traditional IT procedures: a lot of paperwork justification, internal discussions, and waiting. Once a user received a set of virtual servers, they were unlikely to give them back in fear of having to go through the same process again for the next project. Why risk a project delay while waiting for resources? Instead, just hang on to those same virtual instances and leave them running idle. Right sizing was a foreign concept.
With a self-service cloud, IT is a facilitator, not a gate-keeper. Internal customers look to IT for information about resource availability. IT also provides resource cost information to help project owners make decisions about workload placement. IT can then track those requests to better understand which new services to offer and what hardware to add.
Clouds give project owners the ability to rapidly scale up or down according to project requirements. Good news! There’s something called “chargeback” or “showback.” It helps with tracking costs and reporting infrastructure consumption, and also provides a method to control sprawl. By exposing that information, a user will only provision the amount of resources they can justify according to project needs. When the project is completed, the user will return those resources no longer being used in order to lower budget impacts. Of course, don’t expect right sizing if project owners do not have visibility into the resources that were provisioned or the flexibility to rapidly provision future projects.
- Some old concerns are not relevant – Applications designed to run in the cloud are also designed to be restarted. Redundancy and resiliency are built-in. Configuration drift in the operating system doesn’t matter as much anymore. It is not something worthy of being tracked and fretted over. The application is designed to be restarted, and by restarting, the image can be refreshed. Updating OS patches and closing vulnerabilities can take place in a matter of minutes even across large environments. Configuration management is not something needed from your management tool.
The project owner has responsibility for VM right sizing. With a well designed chargeback model, the project owner will pay for the resources being consumed. It is in their best interest to align the instance resource with the project requirements. Otherwise, they end up paying for unused capacity. A good cloud billing solution will not only give resource cost information to the user and the administrator, but also utilization statistics to help make informed decisions about turning underutilized resources off.
- It is a hybrid world – Tradeoffs are everywhere. At Talligent, we have witnessed an OpenStack cloud costing as little as 25% for similar resources provisioned in a public cloud. Yet, there is no expectation that AWS, Azure or Google will go away anytime soon. Instead, use cases for public cloud and use cases for private cloud will become better defined. A public cloud might be the preferred solution for short-term resource needs due to project or geographic requirements. Public clouds are also well suited for risky projects that need market validation before committing resources to building and owning a private cloud.
A good cloud billing solution will give comparative information about the cloud (private and public) options available to the project owner. It is a good opportunity to show project owners that the private cloud can be a cost-effective alternative to public clouds. The cloud billing solution can also better track the public cloud spend on a project by project basis. It can do so alongside the private cloud consumption to help with planning capacity, future year budgets and growth plans. The billing solution can aggregate the data so that executives have a clear view of their department’s IT spend.
- Illusion of infinite capacity – Some tasks, such as capacity planning, have carried over from virtualization management. Providing private cloud services is a delicate balance between having enough resources to support the projects your customers want to run, but not over-buy and have lots of excess capacity. Research has shown that high utilization of hardware resources is critical for a private cloud to be cost competitive with a public cloud. Of course there are other reasons to justify a private cloud (regulatory and security being two of the top reasons), but most companies require them to be cost competitive as well.
Tracking capacity growth requires a historical perspective of consumption trends. It also requires a good understanding of when and how much new capacity will be needed. Plan to add new capacity just as new projects come online. While spending too late can add a large premium to both hardware and labor costs, being too early means tying up precious dollars in underutilized IT assets.
- Bonus #5 – OpenStack is already complex – Don’t make it more so. The OpenStack platform has a lot of moving pieces and it evolves every 6 months with a new major release. Look for a cloud billing partner that is closely involved with the community. A solution that is certified with the key OpenStack distributions will give confidence that support will be given to the latest release and builds. Many distributions offer different packaging and deployment options that might be a consideration as well (i.e. Heat, Fuel, Charms, etc.). This helps simplify the OpenStack deployment and ongoing support.
With 6 core services and new services constantly being added to the OpenStack “Big tent”, the API options are considerable. It is hard to keep up with the metered resource options and correlate them across the services, especially when the Ceilometer services have not been deployed. For true cloud management, a month-end snapshot is not high enough resolution to optimize project resources. A good cloud billing solution should connect with the service APIs to track metered resources and present costs in real-time – no one likes surprises, especially with their bill. Real-time data allows project owners to make adjustments during the month before costs exceed budgets.
For more information about Talligent and our purpose build cloud management solution, please visit www.talligent.com or contact us at firstname.lastname@example.org.