Chuck Dubuque is VP of Product Marketing for Tintri.
Like cutting toenails in transit, some things don’t belong in public. In a similar way, the public cloud might not always be the most appropriate place for all workloads.
This statement has never seemed more relevant than in the aftermath of two latest, very public cloud outages. Both Amazon and Microsoft had to move swiftly to appease users who couldn’t connect to data and apps stored on their AWS and Azure platforms. Many technology companies couldn’t access SaaS-based technologies such as invoicing systems, HR software and A/B testing services. In addition, global multi-platform media and entertainment company Mashable half-jokingly threatened to produce its first printed newspaper. The outages were felt globally, and were swiftly accompanied by the realization that it may not always be wise to put all your eggs in a single cloud basket.
But downtime is not the only reason to avoid putting all your eggs into one cart or public cloud. Yes, outages happen, but for individual companies, cost and performance are the most relevant day-to-day concerns when it comes to public cloud.
Price is More Than Just a Dollar Sign
Many organizations assume that public cloud will cost less than private cloud, since no upfront capital expenditures are required. But there are other costs of public cloud that need to be factored in to decision-making.
Moving applications to public cloud platforms can result in significant migration costs and effort, requiring applications to be recoded, reconfigured, refactored and reintegrated. While a public cloud can scale applications with fluctuating demand, the unexpected cost from unpredictable data growth or the cost of a large-scale cloud deployment can quickly get out of control.
The financial cost and wasted effort combine to provide another challenge. Some workloads and applications are resource-intensive. When this workload sits in a company’s data center, it utilizes free network bandwidth. However, when this same workload is moved to the public cloud it simply runs up a credit card bill. Many companies that have braved the public cloud can share billing horror stories, like surprise monthly bills of $50,000 for network bandwidth usage.
Conversely, the right private cloud can offer predictable costs. Analytics can forecast precise needs for additional capacity and performance to avoid over-provisioning. And when scaling your footprint, automation can optimize the location of every application.
Performance is the Be All and End All
In today’s business climate, IT teams are held accountable for the highest performance standards. As a general rule, the bar for enterprise availability is five-nines, 99.999% availability. To my calculations, with the three hour and 50-minute outage that Amazon experienced last month, it will take them 30 years of flawless performance to achieve five-nines again.
At a more practical level, the public cloud normally works at a consistent level of three-nines or four-nines unbeknown to most organizations that trust the platform with mission critical workloads. And digging a little deeper, applications have to provide their own availability and resiliency in the public cloud, rather than depend on the infrastructure to provide them, which makes it difficult to maintain availability.
It is also important to quantify the impact of downtime to revenue and reputation. How much downtime can each application afford before it impacts the bottom line or causes customers to complain? This should be a primary question when deciding if a workload is right for the public cloud.
Enterprise cloud on the other hand can be the perfect answer for workloads that are more business critical. An enterprise cloud can guarantee availability by eliminating resource conflicts and performance through quality of service controls. Many organizations are discovering that public cloud alternatives can provide the same benefits of greater agility and scalability and lower capital expenditures.
Find the Right Basket for Each Egg
Chances are there is a high degree of variability in your applications. Some are perfectly suited to public cloud. But public cloud may not be the right solution for all applications. Some need the control of cost and performance you can deliver in your data center. That’s why as you construct your cloud strategy, remember not all of your workloads – or eggs – fit in the same basket.
Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Penton.