It is overly simplistic to think that all enterprise workloads should run in public cloud data centers. Today’s digital enterprise journey is much more complicated than that because the strategies involve a hybrid mix of public and private clouds, legacy environments, applications or platform layer, and a broad spectrum of computing resources. Many enterprises are using the services of the public cloud to some extent, and they choose to deploy their workloads on multiple clouds (instead of settling for just one provider).
At ONUG Spring 2019, ONUG Community members discussed how IT leaders can build and run hybrid multi-cloud infrastructure to support their digital transformation initiatives. In this article, we will pay attention to some of the top challenges that leaders need to deal with when choosing a multi-cloud strategy.
Applications need to be platform-agnostic if an enterprise wants to achieve a streamlined multi-cloud deployment. That’s why companies rely on ubiquitous services available with large cloud providers. For example, on Amazon Web Services (AWS), you’ll want to stick with S3, EBS, and EC2 services because they’re the bare-bones infrastructure that you can port to Azure VM. However, many functions behave differently on different platforms, so that’s easier said than done. Some companies decide to build their solutions (for performance reasons) and avoid using specialty services. On the other hand, this requires making all proprietary tools which would otherwise be left to the cloud vendor. Cloud engineers and architects are hard to find, so the downside is also in difficulty in hiring the right multi-cloud personnel.
Relying on a single cloud provider comes with various business risks. Your digital enterprise journey will be dependent on the provider’s proprietary platforms and technologies and pace of innovation. Also, their asymmetric data transfer pricing discourages users from transferring data out of the cloud and incentivizes transferring data into it. These risks don’t exist with a hybrid multi-cloud solution. You will avoid vendor lock-in, make sure that you are not dependent on the technology of one provider, and will strengthen your position when negotiating terms and prices with the cloud provider.
Intra-cloud data transfer costs (when operating with one cloud provider) are often free or very low. But when it comes to transferring data to another cloud provider, it’s a different case. These transfers can be pricey, while the amount of data transfers grow. Thus, operating in a multi-cloud environment can cost you a lot at the end of the month.
Cloud outage is another connectivity consideration. Traffic inside the cloud usually works, even during massive outages. It allows your virtual machines to talk to each other. However, that’s not the case when it comes to inter-cloud connectivity. In case you have to sync the service between different clouds – pay for better connectivity, utilize asynchronous systems as much as you can, and never let a cloud-specific service become a point of failure.
Today, enterprises are looking for a mix of multi-cloud and hybrid infrastructure. That has led them to implement cloud technology in a way that wasn’t anticipated in the past. One of the most critical functions remains to be on-premises deployments because an enterprise cannot just give up on their existing infrastructure investments. Technologies such as ML and AI could help solve the complexity issue. Both present a massive potential, and by gathering substantial amounts of data and pulling them together will help us see the patterns more clearly and optimize for the future. Simplifying the complexity of hybrid and multi-cloud can be achieved with a combination of automation and AI.
Another critical concern for today’s enterprise is the skills gap. Who in the enterprise has the expertise and knowledge to oversee all of the functions involved with hybrid and multi-cloud deployments? They need to understand both the technology and economic implications (such as budgets and forecasting). Typical IT admins or directors don’t have those skills, because there’s the specialist equipment to be managed. It is where enterprises need to look for help from a more experienced partner in this complex and rapidly changing environment – data scientists as well as experts in AI and data analysis. These are the people that can help drive the digital enterprise journey by helping to pull vertical sectors together.
Hybrid and multi-cloud are two concepts that are a part of a larger picture, not mutually exclusive alternatives. These architectures comprise a hybrid of pay-per-use public cloud services (from multiple cloud providers) and owned enterprise resources. The hybrid multi-cloud solution can support an integrated strategy utilizing centralized, distributed, and dispersed storage and processing. By balancing these elements, companies can maximize agility, improve user experience, reduce cost, enhance reliability, accelerate time-to-value or time-to-market, etc.