Enterprises working with cloud applications are embracing hybrid and multi-cloud resource provisioning more frequently than in past years. That’s because multi-cloud gives developers a great choice of using AWS, Azure or Google resources (e.g., compute, storage, networking, etc.) when they need them.
However, this can lead to two problems:
- You typically need to be an AWS or Azure expert to get those resources (because the AWS or Azure portal is complex).
- Without guardrails around cost and security in place, access to those resources is non-controlled, which can lead to compliance issues and/or cost overruns.
The remedy to these issues is blueprints. Blueprints are templates for publishing a catalog item that IT admin creates so that a cloud user can request the required resource stack. These blueprints are cloud agnostic and can simplify the provisioning process. They can also be used for Day 2 services, such as scaling, deleting, etc. to provide more agility to the users.
Basically they are IT's answer in delivering the simplicity/agility that developers want (through a self-service portal), while ensuring that IT can add appropriate guardrails in the consumption of those resources (e.g., making sure cloud resources shuts down when not in use; making sure cloud resources only run during certain times of day, etc.) As the deployment of hybrid clouds continues to proliferate, we are seeing more and more cloud management providers deliver blueprints as a way to balance agility with control.
Industry information for this eWEEK Data Points article was supplied by Grant Ho, chief marketing officer at CloudBolt Software. Here are five data points Ho believes companies should consider when utilizing blueprints for hybrid cloud management.
Data Point No. 1: Simple or Complex?
Blueprints can be as simple as single servers (e.g., AWS EC2, S3 buckets,) or as complex as multi-tier application stacks (e.g., OS / database / web server). Companies need to decide which setup for deploying blueprints will work best for their needs. The best way to decide how sophisticated the blueprints need to be is to first determine how many cloud services need to be managed. Then the company needs to assess their IT staff’s expertise in the various services. The blueprints should then be gauged to bridge the skills gaps between the services.
Data Point No. 2: Create Modular Actions
When designing a blueprint, it’s important to modularize your approach as much as possible so that what you use as an “action,” such as sending an approval request to an IT admin or getting the underlying password, can be reused in multiple environments.
Data Point No. 3: Control Configuration Settings
As you determine what resources need to be totally standardized for end users, with size restrictions for any compute settings such as CPU, memory or disk size, you can set it in the blueprint. On the other hand, you can make them available to set at runtime for the blueprint to use when deploying. You could even use orchestration settings that pick the best environment based on input from the end user.
Data Point No. 4: Show Associated Costs
Being able to display costs that you set for specific resources or to retrieve the associated costs from public cloud providers for ephemeral use of specific resources helps end-users determine a best fit based on their budgets. This level of control with associated costs helps IT admins control and govern the provisioning of multiple requests for multiple modes of delivery.
Data Point No. 5: Preserve Tribal Knowledge
One of the biggest issues facing IT departments is retaining top talent. When people with specialized knowledge of services and systems leave an organization, either through a new job opportunity or due to illness, it can create a skills gap within the organization. By utilizing blueprints, enterprises capture and retain that knowledge so that other staff members can still adequately handle tasks even though they are not experts in that particular skill set.
If you have a suggestion for an eWEEK Data Points article, email email@example.com.