One of the more interesting debates going on in the cloud computing area is whether CIOs should select multiple vendors for their cloud strategy.
In the pre-cloud era, there was no debate: IT systems were always composed of hardware, software, storage, and networking equipment from different vendors. At that time, these choices were powered by each company’s specific needs as well as any pricing considerations from individual vendors.
Another key advantage of choosing components from different manufacturers was the element of risk mitigation brought in by which components could be switched out in case of failure or nonperformance from one set of components. So why not do the same with cloud systems? Here, we look at the multiple vendor cloud strategy.
Costs and alternatives: Key points to consider
The notion of one cloud platform that provides high scalability and security at low cost and without vendor lock in remains somewhat of an idealization. Invariably, in an attempt to reach for this idealization, organizations are likely to become multivendor and multiprovider.
One of the key points to consider is that unlike the earlier days of the cloud, alternatives are very much available in the form of AWS, Microsoft Azure, and Google Cloud. AWS has become No. 1 in cloud computing, but Azure has grown by providing a transition path from Office to Azure via Office 365, while tossing in several volume discounts at companies along the way. Meanwhile, Google Cloud has landed big-name clients like Spotify and Snap.
Besides, the fact of the matter is that different departments in an organization often have a tendency to have services developed on different platforms, which means that cloud management teams necessarily need to be in a position to support environments hosted by multiple providers.
Technology and regulatory aspects
At the core of it all lies the question: What is the right cloud service for my application? If you are running apps that are largely .Net, MS Azure would be a judicious choice for your deployment. Whereas if your apps are dominated by Java and Linux, AWS remains the most suitable choice. This is not like Transformers movies, which are made for everyone, but more like a movie such as “Jurassic World,” which was apparently only created for children and not people who wanted to see something different than what was shown in “Jurassic Park.” But I digress.
Other considerations relate to how each cloud service provides compliance with standard industry regulations like HIPAA for health care and Dodd-Frank for the financial sector. While it is true that all leading cloud service providers will give you compliance with industry regulations, some of them are more adept at this, even at a minute level.
Lock in: Should it be a concern?
Lock in is not as bad a word as it seems to be. And don’t confuse this with Lockdown from “Transformers 4”! (But I digress again.) Moreover, whenever your organization purchases a solution from a company, in some ways, you are locked into that solution in the sense that there is a cost associated with migrating out of that solution.
It is not true that you absolutely cannot migrate out of that solution. It is no different from the kind of lock in you face for your email system or the CRM solution you have installed. As such, it is not hard to move loads between providers unless you are taking advantage of many of the specialized services offered by the vendor such as domain name, relational database, and load-balancing services.
It is critical that CIOs arrive at the right balance between the desire for innovative, high-performing apps with the desire to minimize cost and achieve high efficiency, since both priorities can work at cross purposes with each other.
The fact of the matter is that in order to be able to run high-performing apps, you will invariably have to support cloud configurations that are expensive. Similarly, when you are trying to minimize the cost, it often happens that the best possible cloud configuration cannot be built and supported.
It is highly recommended that you start small with a couple of cloud partners and spend time to build the expertise in working with these few providers. It is always a sagacious idea to test a cloud-based solution before actually signing an agreement.
This gives you enough time to assess your comfort level with the vendor before going ahead and actually signing the agreement. Also, it is a sterling idea to choose the vendor based on the design and architecture of the application – different vendors have different strength areas in terms of technical expertise.
As long as you remember that not every application needs to be able to provide 99.999 percent availability with the ability to run on two different nodes simultaneously, you will be in decent shape. Most multiprovider architecture involves these kinds of setups.
Assess compute and workload management services
The services related to workload and compute management include cluster management, orchestration, and configuration tools. Virtual machines being the building block of any compute service, it is important to understand the differences between various VMs that are available. The OS, memory, processing power, and network optimization features are all part of a VM’s specifications.
Orchestration and configuration tools
Features related to orchestration allow administrators to automate code deployment and define code infrastructure. It is particularly paramount for those organizations that need scaling or those that run across multiple clouds. Orchestration and configuration tools such as Puppet and Chef are awesome for multicloud use.
Vendor specific tools including AWS CloudFormation are also possible options. While third-party tools offer cross-cloud flexibility and permit the moving of workloads across clouds, vendor specific tools would be custom built to work within the specific vendor’s cloud.
Unified administration and management
When deploying hybrid cloud environments, one mistake that organizations can make is that too much complexity can be introduced into the management stack in terms of numerous software tools overlaying automation on native management toolsets.
Unified administration and management is a critical need in a hybrid cloud environment. It is that kind of unified management that can provide the agility and cost advantages of on-premises and blended cloud resources. Tools like VMware can be extremely useful when needing to deal with a combination of multiple deployments. VMware can be used to manage off-premises resources as well.
Photo credit: Pixabay