BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Cloud providers don't offer uniform services. What's available from one won't always be accessible from another. This creates the possibility of mix-and-match scenarios -- a business can have applications, and even data, that span two or more clouds.
Blending public and private cloud services from more than one provider helps a business avoid vendor lock-in, deepens the pool of available services and allows the company to maintain control of its most sensitive data. The downside to using multiple clouds is the added complexity, especially to administer applications, APIs and configuration management.
Cloud customers need to decide if those situations warrant the investment. If the answer is yes, determine which providers best deliver particular services. Cloud service brokers (CSBs) manage complexity and arrange services as needed. But the downside is CSBs are expensive and can be confusing to deal with in their own right.
You can't toss tools at this one
Businesses typically target multiple clouds for a few core reasons:
- A single cloud locks the business into a single set of cloud services, which will be difficult -- if not impossible -- to decouple from later.
- Cloud services evolve quickly, and companies want to pick the best public cloud provider, but don't know who will ultimately rule the market. Amazon Web Services (AWS), for example, pushes out dozens of cloud services a month, and it changes them as needed to meet demands and expectations of its users. Microsoft and Google take similar approaches.
- Finally, businesses believe that multicloud tool sets, including cloud management platforms, or CMPs and CSBs, will give them a single platform to use, easing complexities and platform differences.
The truth is that tools, including CSBs and CMPs, won't save a company from multicloud's complexities and issues. Businesses that use multiple clouds quickly discover that mixing security services between two public clouds is like trying to shove a VHS tape into a cassette deck.
Public cloud providers' respective public cloud-based security options, such as identity and access management, are not built to be compatible. Instead, an IT staff needs to create or purchase -- and then maintain -- a security product that spans two or more clouds, which increases complexity and cost.
CMP and CSB tools provide some relief for underlying native cloud services, but the cost to implement these tools often outweighs any value gained. Moreover, as cloud computing evolves, the tools likely won't keep up. IT admins will bypass the tools to deal with the native clouds more often than anyone expects.
Finally, cloud providers will likely purchase these tools themselves -- and you can guess which cloud they will work best with. Cloud providers have done this before, and they essentially put the tool sets out to pasture. That's a problem if your business built its multi-platform strategy around that tool.
Some organizations also consider serverless cloud computing. Serverless is a bit misleading; there is a server, but -- unlike traditional public cloud servers -- the business doesn't have to provision the server to run applications, such as storage and compute.
Serverless comes on the heels of containers, which offer cloud portability as a benefit. However, each public cloud provider has a different approach to support containers. It doesn't matter if you run services on AWS Lambda, Azure Functions or Google Cloud Functions -- you'll be locked into those systems since each uses its own programming language set, structure, data connectivity, security and so on.
Serverless might be a step too far for many businesses. The technology itself means that companies must decide how to form, implement and manage multiple clouds. It might be easier to choose one -- either AWS, Azure or Google.
Serverless technology is compelling, as is the fact that you can count on more technology shifts down the road. So ask yourself: Will you go native cloud from the jump? Or will your business return to native cloud after it spends a few years and a few million dollars? Other organizations are considering these questions right now.
Kubernetes, Docker push multicloud forward
Red Hat looks to break through multicloud barrier
Time to evolve a load balancing strategy