With Rising Costs and Vendor Lock-Ins, Is a Cloud Exodus in the Making?
Public cloud hosting has become the unofficial business standard, but it comes with a steep price.
Public cloud hosting has become the unofficial standard for businesses these days, even more so because of Covid-19. About nine out of ten businesses accelerated their cloud use amid the pandemic, according to a recent Flexera survey, and that trend began long before the virus. Cloud deployment and its variants, such as hybrid and multi-cloud, have long been a high-tech staple since they empower businesses to scale up quickly and with minimum friction.
With greater cloud usage, however, comes a greater price tag. Infrastructure-as-a-service solutions will haul in over $106 billion in 2021, Gartner predicts, up from some $82 billion in 2020 — the largest increase across end-user spending on public cloud. The figure is indeed significant to businesses, with 77% of companies stating they were surprised with their cloud expenses in a 2021 survey.
Even worse is the fact that some of that money is not being spent wisely. Some estimates put the amount of money companies will spend on idle resources at over $15 billion this year. The assessment zooms in on the capacities designated to run testing or development deployments, which are not needed outside working hours but were kept online 24/7, cranking the cloud expenses up.
Despite the preeminence of cloud architecture, we’ve seen numerous companies drop public cloud, whether partially or in full, in favor of setting up their own infrastructure with reasonable success. Dropbox is the highest-profile example, saving some $75 million over two years after bidding farewell to the Amazon cloud. And it’s not alone in this push.
Like anything else, cloud architecture has its own flaws. Public cloud expenditures make dents in software companies’ bottom line and market capitalization, a16z argued in a recent research. Other critics point at the fact that a major share of the public cloud market is held by only three giants — Amazon, Microsoft and Google — and lambast the public cloud costs as unsustainable. Another problem that could make companies consider running their own data centers is the ever-present issue of vendor lock-in: Once you move your business to the cloud, getting out would be costlier and costlier.
Managing costs and expectations
To a degree, the problem simply comes down to managing the expectations. Migrating to a cloud comes with its own costs, especially early on, as a company’s developers rework the solutions for their new runtime environments. From there, the ruthless logic of business would kick in: After all, it is in the cloud provider’s best interest to have customers pay more over time, especially once the costs of moving to another cloud or architecture are sufficient.
It should be noted that businesses are not powerless in all of this, with a whole range of tools available to them to streamline their cloud budgets. Some of those are provided by vendors themselves: The AWS Cost Explorer, for example, is a good starting point to estimate the effectiveness of the deployment on the Amazon Cloud. Going forward though, companies tend to adopt more third-party tools to get a better picture of their spending.
Keeping track of the expenses is especially important for companies relying on a multi-cloud or hybrid architecture involving several vendors. Such networks are more complex, and a bird’s eye view of the operation makes it easier for FinOps teams to keep the expenses streamlined.
This problem can also be solved by working with vendors that aggregate a variety of cloud solutions and enable businesses to distribute their operation between different hosting services, optimizing every aspect of the deployment. This effectively provides a multi-cloud architecture delivered via a single platform, which simplifies the migration and management.
Another important way to slash cloud expenses is by using AI-enabled tools to automate cloud management. Overall, automation is the name of the game with cloud migration, as such smart algorithms allow companies to focus their own IT desks and FinOps teams on non-routine, strategic tasks as AI takes care of scaling the virtual-machine fleet in line with the business needs.
Time to repatriate?
It is true that today’s public cloud market is not without its flaws. And yet, it is hard to imagine companies rolling their operations back to their own premises. Those awed at their own public cloud spending would be just as surprised by the repatriation costs too, after all. The process would include buying their own servers, renting premises, running maintenance and all the associated integration costs along the way.
Granted, there may well be a stage at which it would make sense for a company to pull out of any and all third-party clouds. But reaching this stage is not easy. Developing the software and hardware needed for the transition will take a lot of time and resources, and having to maintain and scale your own data centers also comes at a price.
For those with neither the time nor the money to fuel such an endeavor, the public cloud remains a lucrative option. Early-stage tech businesses and startups use public cloud as an easy and accessible way to deploy their solution and start growing at once. For established businesses, including giants like Apple, public cloud is a convenient addition to their own private clouds that can be used to host specific products or provide extra capacities for peak-load hours. Thus, as public cloud use looks set to only grow more and more omnipresent, the right thing for most businesses to focus on is building up their FinOps capabilities and getting better at managing their footprint in the clouds.
Entrepreneur Leadership Network Contributor