When I started my career in information technology, I worked for a large insurance company in Omaha, Nebraska. At the time, they exclusively used Lotus Notes, an IBM product. Even as Microsoft Outlook gained popularity and functionality, the cost of changing email clients was insurmountable, so the company continued using Lotus Notes for many years.
Thankfully, being trapped on a client application is far less common nowadays because modern apps are more flexible and frequently upgraded without user or administrative intervention. However, organizations can still explore having multiple solution providers to help avoid vendor lock-in, provide redundancy and resilience, and lower costs. For example, implementing multi-cloud management is becoming a popular option but can also come with some potential challenges. In this post, I’ll explain the fundamentals of multi-cloud, challenges to consider, and how the
What is Multi-Cloud?
A new architecture topology currently gaining momentum is multi-cloud or hybrid cloud. This is an architectural design that allows organizations to span multiple private or public cloud service providers simultaneously. A popular business model is multi-cloud through Infrastructure as a Service (IaaS), illustrated in Figure 1.
Figure 1: Three sample cloud services integrated into a multi-cloud architecture
According to Gartner, this model is implemented by 75% of organizations, up 26% from only 49% five years prior. This approach allows organizations to utilize technology offered by each cloud vendor while potentially offering a robust and stable infrastructure. A multi-cloud approach is usually implemented in a tightly integrated manner tosynchronize data back and forth seamlessly. data back and forth seamlessly.
What are Common Problems With Managing Multiple Clouds? Benefits vs. Challenges Examples
When done correctly, the return on investment of a hybrid cloud topology can be immense. But the orchestration of such an environment comes with its own set of challenges that must be addressed. Let’s compare a few key concerns and benefits of multi-cloud management worth mentioning.
Increased architectural complexity
One of the benefits of sticking to a single cloud provider is that organizations can become very fluent in how that topology functions. Engineers understand how to expand the environment and mold it to meet business requirements. Bringing in other technologies increases the solution’s complexity.
The additional complexity of multi-cloud management usually requires additional time and expenses to ensure the solution is implemented correctly. The complexity most likely spans several areas of technology, such as networking, automation, data services, recoverability, and security. Moreover, this evolution could result in expended costs due to having to retain individuals who have specific knowledge about each specific cloud provider.
Although individual products, such as Microsoft SQL Server, would operate the same as before, an SQL Server database administrator who knows and understands the Microsoft Azure ecosystem might not uniquely understand, for instance, the Amazon Web Services (AWS) environment. Each cloud provider brings its own unique set of issues that must be addressed. Specific knowledge about each one is just one facet.
Some of this complexity could be mitigated by using
practices such as DevOps and management tools such as Kubernetes, which abstract some of nuanced differences between cloud platforms. DevOps can facilitate an agnostic workflow to any public cloud provider. Kubernetes allows a more seamless experience regardless of the underlying cloud provider. The pods deployed would continue in the same manner to provide API or application services whether in Microsoft Azure, AWS, or Google Cloud Platform (GCP).
While such general tools can simplify application deployment, the infrastructure team needs to be aware of networking and security differences as they deploy Kubernetes services on each of those clouds.
Disaster recovery planning
In any dealings with a cloud provider, reliability and recovery should be paramount. Quite often, in my experience, organizations that invest in cloud technology believe that it will solve all the companies’ pain points. They seek a “set it and forget it” type of solution, which is far from achievable.
One commonly overlooked issue is disaster recovery. Disasters happen even to cloud providers, so organizations must plan to recover when (not if) that occurs. Understanding the failure modes of each cloud provider and how they might affect your application stack is crucial to building reliable applications. Due to the increased complexity of multiple cloud providers in the mix, this problem should be at the top of the priority list from the start.
Recovering from a disaster across multiple cloud providers can prove to be difficult. Often, each cloud provider has a proprietary method to enact a recovery process. This increased complexity could further inhibit your predefined recovery point objective (RPO) and recovery time objective (RTO). Compromised RPO and RTO metrics could lead to violations of service level agreements down the line.
Consistent reliability is critical for any deployment, regardless of the medium you’re deploying into. A multi-cloud architecture can be beneficial to organizations by offering multiple layers of redundancy. This setup allows organizations to better control the risk of failure across multiple silos. For example, if a single cloud provider is having an issue like a domain name service (DNS) failure, the other cloud providers will not.
Failure scenarios also raise questions around how you handle multi-cloud failover—for instance, which cloud you “home” DNS records in—and how you handle the resulting transition. Large scale web companies do this by hosting those solutions outside of the networks that host their applications. If you need extremely high levels of availability, you may need to consider this extra layer of redundancy. This is also true when it comes to physical redundancy, such as data centers. Each cloud provider handles its own physical demands, such as electrical, water, cooling, etc., thus providing another layer of redundancy. While disaster recovery must be considered even in a multi-cloud infrastructure, you should take advantage of the natural redundancy offered by such topology.
Cloud sprawl and costs
Sprawl can manifest in any aspect of an organization, from resources to
tool sprawl. Sprawl is manifested as an abundance of things, in this case cloud resources, that have become unwieldy and even potentially unmanageable without intervention.
In previous years, a common problem typically manifested itself as server sprawl, where servers would be created, implemented, and subsequently forgotten. The process would repeat itself, and before long an organization would have ownership in a large server farm that it didn’t even know existed. Cloud ecosystems can also become unwieldly and cumbersome.
Thanks to the sheer ease of quickly creating cloud resources like virtual machines, organizations are more susceptible to sprawl. This self-service model of provisioning cloud resources within minutes leaves a shorter retention time for organizations to do something with it. This kind of overuse can also be manifested in a common horror story about a cloud resource being created and forgotten, causing a substantial bill for the organization and can potentially reduce the bottom line of any organization. Proper cost management is crucial.
Managing security policies
Another major requirement in this architectural design is managing security policies. Given cyberattacks are becoming more prevalent, a multi-cloud ecosystem can widen the attack vector for a bad actor to potentially compromise.
Organizations must provide a higher level of due diligence to secure all their boundaries, now across different cloud platforms. In a multi-cloud environment, this can be challenging due to the different kinds of endpoints that exist in each cloud.
Finding the right multi-cloud management tool
When you expand any ecosystem to a more complex and potentially tightly coupled environment, consistent and reliable multi-cloud management can become difficult. While each cloud provider may provide its own monitoring tools, these tools often lack any holistic overview outside their own confines and manually monitoring or using disparate tools to piece together performance insights can often lead to more frustrations than resolutions.
Instead of just monitoring. make sure you can observe what’s really going on in your multi-cloud environment. Observability insights can help provide more comprehensive visibility and actionable intelligence to help bring focus and clarity to complex environments.
SolarWinds Observability Self-Hosted (formerly known as Hybrid Cloud Observability) is built to centralize observability insights from your IT infrastructure and services in a single pane of glass for your entire ecosystem whether on premises or in the cloud. This increased visibility across your IT infrastructure and services can allow for easier management, reduced resource sprawl, better cost management, and overall availability.
SolarWinds Observability Self-Hosted (formerly known as Hybrid Cloud Observability) can also help you ensure proper multi-cloud management optimization at different stages within the cloud deployment lifecycle and other digital transformation strategies, such as through a cloud migration from on-premises as you examine future expansions and integrations into a multi-cloud infrastructure. You can learn how one national communications provider operating a hybrid IT environment with a multi-cloud infrastructure was able to save millions in recurring costs with SolarWinds Observability Self-Hosted (formerly known as Hybrid Cloud Observability)
here.
