In a multi cloud environment, a company uses several cloud services from different cloud providers simultaneously by connecting the offerings of several providers and enabling different SaaS and PaaS applications to communicate with each other. For example, a company simultaneously uses public cloud services from Oracle, Microsoft Azure and AWS for the same purpose.
The multi cloud is actually an extension of the hybrid cloud, but offers far more possibilities than the hybrid form. Multi cloud is especially relevant for multinational and large companies with many different applications and expect the full cloud benefits such as flexibility and choice.
Whether the multi cloud environment became unintentionally as part of the historical cloud journey or whether this is a strategic decision such as fear of cloud vendor lock-in, it is always a unique decision of company to determine what to do first. For some areas a single product, i.e. CRM, drives which cloud platform to select, in another area a company sees the advantage to move their whole data warehouse into the cloud, in other cases a company sees an advantage to move their data center into a private or public cloud.
Another question is the “make” or “buy” approach. Do we have enough in house skills to build a private cloud or do we need to hire externals or a consulting company to make it happened? Can we select a single public cloud service provider to support all workloads or only specific ones?
Most companies which run some of their workloads in the cloud are now facing the decisive phase of cloud use, which deeply affects the infrastructure and contains bigger risks if the wrong decisions are taken.
The questions multi cloud users need to address are:
- How can we keep costs under control in those large digital transformation projects and at the same time ensuring the long term performance of our IT?
- What are the key drivers for powerful data management when following a multi cloud approach?
- How to address availability, performance and service level agreements in the cloud to be successful?
Today, companies following the multi cloud approach use half of their cloud services for Infrastructure (IaaS), followed by storage, Platforms (PaaS), basic applications (i.e. SharePoint and other collaboration applications), application development and core applications (i.e. cloud ERP).
There are many multi cloud possibilities especially in the Software (SaaS) area when the majority of companies using Infrastructure (IaaS) services.
Multi cloud example
For example, a company in the direct sales decides for the multi cloud first approach and takes cloud advantages by virtualizing and standardizing client relationships and mobile workplaces. A switch to VOIP provides more flexibility to the workforce and the call center.
Considering the “make” or “buy” approach, the IT Strategy can look as followed when using a Business and IT alignment project methodology:
- Cloud First approach
- Best of Breed approach at domain level
- Mapping of the IT delivery teams to the domain
- Use “make” approach for USP processes and “buy” approach for standard processes
- Start with frontend systems such as POS, Call Center, new customer acquisition, B2B and B2C online business
- Next level applications: CRM, Order Management, Logistics and ERP
- Further moves: Data Warehouse (DWH), Reporting, ML, KI – cloud analytics
Regarding DWH, when using more than one cloud service provider, the data is scattered across multiple public clouds. Cloud service providers such as Google offer multi cloud analytics solutions such as Big Query Omni. This tool lets you access and analyze data across the Google Cloud, Amazon Web Services (AWS) and Azure. So the customer can keep their big query user interface or standard SQL queries and will be able to query the data from different clouds without managing the different and underlying infrastructures.
For each wave, the right master plans and processes must be aligned between IT and business. What are the new technologies to take advantage of? Explain to business what are the risks and benefits when moving to the cloud, address questions such as data security and data migration.
Multi Cloud Risks
Having made first positive experiences with moving workloads to the cloud, companies consider putting more workloads into the cloud and with different cloud service providers as it appeals to mitigate risks, reduce vendor lock-in risks and getting benefits from cloud provider’s expertise.
However, there are risks with the multi cloud models which need to be considered for
- portability costs
- licensing costs
- price arbitrage
- own data centers to move to the multi cloud
- increased number of applications
- data protection / compliance
- performance
- data storage
- data management
- location
- risk mitigation because of diversification
- degree of lock in
- portability and value added services
- automation capability
- tools evaluation and selection
- availability and lack of skills. In multi cloud, you need cross platform skills
- redundant, combined or autonomous multi clouds
Companies need to manage potential conflict of interest with secondary cloud providers and evaluate alternatives such as single provider in a single region vs single provider in different regions.
In order to have a sustained multi cloud approach, companies need to check the cloud provider interoperability, enhanced inter cloud connectivity, API standardization and the maturity of the Kubernetes ecosystem.
Price, Costs and Licensing Risks
Multi cloud means using different service providers for the same activity, for example provisioning machines from different vendors. From a price, costs and licensing risks perspective, many companies chose unintentionally different vendors or different cloud environments, where business decision makers became front runners and overruled central IT strategies. As a consequence, costs and complexity are increased and it takes more skills and tools to support multi cloud environments. However, there are more companies operating in a multi cloud environment intentionally instead unintentionally, hence, the multi cloud decision is wanted.
Companies continue to shift costs from on premises to the cloud with the goal that benefits and costs will predominate the risks when using the cloud. Multi cloud is more cost effective when it comes to set up, running, maintenance and support of the services provided from the vendor. The initial thoughts regarding cloud billing is about billing per usage respectively pay per use subscription. So as a conclusion you can say it eliminates the costs of servers, network, maintenance, patching and upgrades since the services you subscribe are internet based. However, it really depends on services, capacity and space a company requires. The total cost of ownership may become higher if not considering the entire life cycle of the cloud services purchased compared to the lower upfront and migration costs.
A lift and shift approach for moving applications to the multi cloud is not recommended. Cost will occur when the application came from legacy systems or is running on premises. As a prerequisite, the application needs to be checked how well it fits into a distributed multi cloud environment and testing is required whether software code needs to be rewritten or configured in order to function properly in the cloud.
Prices for Central Processing Units (CPU’s), computing nodes, storage and networking can end up in nasty surprises if not studied carefully. Transferring data into the cloud is much less expensive than taking data out of the cloud. Companies should pay attention to any data withdrawal charges.
Since cloud service providers are building more and more data centers around the world to serve more local demands and to meet data security and legal requirements, companies should pay attention to geographical price differences because prices are not necessarily cut if a customer is not serving a specific region or does not need an offered location from the cloud service provider.
Another price estimate option is to calculate the damage costs when the application is not used at all. In that case and if the impact is low, there may be no value to move the application to the cloud.
The cost of the multi cloud should be transparent compared to traditional on premises servers and hardware, there is a monthly subscription fee with no extra cost for updates or customization. However, you still have the security, management, and integration overhead of running across multiple clouds.