With all the different cloud computing services available, it can be difficult to choose a provider. One option may be more appropriate for some of your needs, but it may not cut it for every workload or type of operation. The good news is you don’t have to settle for a single cloud. The multi-cloud, a combination of several computing environments, and it is different from a hybrid IT model.
This option is becoming increasingly popular. However, multi-cloud environments are complex and are thus not simple to manage. There are a number of important considerations you should be aware of before you attempt to move to a multi-cloud deployment.
This is a series of articles about hybrid cloud.
In this article:
- What is multi-cloud
- Using hybrid cloud management tools for a multi-cloud
- Cost optimization
- Reliability and disaster recovery
- Containerized and serverless workloads
- Security and compliance
- Multi-cloud storage with Cloudian
This article is part of a series on Hybrid IT.
What Is Multi-Cloud?
A multi-cloud is a system composed of more than one public or private cloud service. Multi-cloud is similar to the concept of a hybrid cloud, but with an important difference: in a multi-cloud, there are several cloud environments but they could be completely separate, or run different workloads. In a hybrid cloud, there are more than one cloud environments, which are tightly integrated and allow seamless transfer of data and workloads between clouds.
- Multi-cloud is very widely adopted—according to IBM, as of 2018, 85% of enterprises operated a multi-cloud. There are several reasons multi-clouds have become so common:
- Cost and lock-in—the classic use case for multi-cloud is to give organizations leverage and avoid lock into a specific cloud provider. If you run on, and are integrated with, several clouds, you can move workloads to the cloud provider that offers the best value for money.
- Shadow IT—organizations increasingly see software deployed by teams without coordination or permission from the central IT organization. Many of these shadow deployments occur on public clouds, which inadvertently creates a multi-cloud.
- Variable cloud capabilities—each public cloud provider or private cloud platform provides different features, and specific workloads or projects within the organization may not be satisfied with a specific platform’s offerings. Making everyone happy with just one cloud provider or private cloud technology proves difficult over time.
- Geography—each public cloud provider has data centers in different geographical locations. If you need proximity to specific branch offices or end-users, you may find it optimal to run workloads on specific regions deployed by several cloud providers.
- Business continuity—having multiple clouds can improve your resiliency to outages or configuration issues that are specific to any one cloud provider.
In the rest of this article, we’ll offer several critical considerations you must be aware of, as you plan your multi cloud management strategy.
Consideration #1: Using Hybrid Cloud Management Tools for a Multi-Cloud
There are few tools dedicated to the special case of multi-cloud management. Multi-cloud is a special case of hybrid cloud, and therefore Hybrid Cloud Management (HCM) tools are the best choice to help you manage a multi-cloud, provided they support all the public clouds you plan to use.
Examples of popular HCM tools are RightScale, Scalr, HyperGrid, as well as open-source options like OpenStack and Apache CloudStack. These tools typically offer the following capabilities:
- Service aggregation and cost management
- Self service for end users
- Release and deployment orchestration
- Workload and cost analytics across clouds
- Integrations and APIs with enterprise and dev tooling
- Platform as a Service (PaaS) allowing users to run infrastructure as a managed service
- Automated workload migration between public and private cloud environments
Consideration #2: Cost Optimization
Use a variety of strategies to optimize the cost of your multi cloud:
- Shop around for cloud providers and negotiate with vendors.
- Consider using cloud services brokers (CSB), consultants who perform manual or automatic evaluation of cloud resources and suggest optimal strategies.
- Some CSBs offer integration or management services that can further reduce costs.
- Use cloud optimization tools—these are offered at a basic level by each cloud provider, but you’ll benefit from using a cross-cloud cost optimization technology
- Build cloud-native applications whenever possible—this will let you make optimal use of resources
Consideration #3: Reliability and Disaster Recovery
A multi-cloud strategy is often a way to achieve higher reliability and business continuity. At the same time, you need to think holistically about the reliability of the multi-cloud deployment itself. Understand how cloud-native resources should behave in case of disaster, and create backup and migration paths to enable this behavior.
Don’t ignore backup and archiving—cloud-based resources should be backed up just like on-premise resources. It is a good idea to leverage the multi-cloud by backing up resources from one cloud to another cloud provider. In some cases, standards or regulations may require you to backup or store copies on-premises or in a specific geographical location.
Consideration #4: Containerized and Serverless Workloads
It is increasingly common for applications to run as containerized workloads, managed using container orchestration tools like Kubernetes. Containers make it easy to move application instances between clouds in a predictable way, but at the same time add a layer of complexity in terms of security, networking and resource provisioning. Even if some of your applications are not containerized at this point, ensure your multi-cloud is container friendly.
Another growing deployment option is serverless. Serverless workloads are complex to manage across clouds because they typically depend on proprietary infrastructure management systems, like AWS Lambda. Find ways to “translate” serverless functions from one cloud provider to another, or at the least, define policies that allow data transfer and interfaces between serverless apps running on different clouds.
Consideration #5: Security and Compliance
Be aware of standards and regulations that affect your multi-cloud and plan for compliance. Control access to data and applications in a unified way across all multi-cloud workloads – a common way to achieve this is Identity and Access Management (IAM) solutions.
Encrypt data, both data at rest and data in transit, and ensure you have a robust strategy for managing keys and certificates across the multi-cloud. Ensure your security team treats the entire multi-cloud as one security perimeter, with common tooling and strategies for securing systems in each public cloud.
Multi-Cloud Storage with Cloudian
Hybrid and multi-cloud infrastructures are not always simple to manage, but it’s easy with the help of Cloudian. Cloudian’s HyperStore is a scalable, on premise object storage platform with a 100% S3 native API. With HyperStore, you can easily manage your data in both public and private storage, and you can integrate it with a variety of cloud and third-party services, including migration services.
Cloudian allows you to easily manage stored data using bucket-level policies, such as backup scheduling and lifecycle. You can automate erasure coding and replication, as well as automatically verify and encrypt data. You can use custom metadata tags facilitate search and analytics.
HyperStore will help you store your data securely and efficiently, so you can focus on your actual work and get the most out of your multi-cloud configuration.
This is a series of articles about hybrid cloud.
