In 2020, federal spending on cloud computing topped some $6.6 billion, up from $6.1 billion in 2019, according to an analysis conducted by Bloomberg. From civilian agencies like the Department of Health and Human Services to the center of U.S. military intelligence at the Pentagon, 2020 witnessed the largest dollar amount invested by the federal government into cloud computing as an increasing number of agencies continue migrating their data and services to the cloud.
With the goal of bringing added strength to security measures and long-term value in perpetuity to government organizations offering key services to citizens and federal personnel alike, this increase in spending on cloud computing signals an additional hurdle to overcome: the interruption of critical services in the event of cloud outages.
Here’s the thing: cloud outages are nothing new. Our innovations — much like ourselves — are imperfect. The cloud is no exception to this. In fact, the services offered by the ‘Big 3’ cloud providers have commonly experienced outages in the past due to a number of factors like routine maintenance, equipment failure, and even malicious cyber attacks.
With more government agencies and organizations continuing to move to the cloud, any outages on the end of cloud providers could cause widespread disruption to critical government services. Since outages could put valuable confidential information — or even national security — at risk, we must enact interoperability of these services across multiple cloud providers, allowing them to continue in the event of an outage.
Sleeping Past the Wake-up Call
Just last November, Amazon Web Services (AWS) experienced a major outage that disrupted service to Adobe, New York City’s Metropolitan Transport Authority, the Washington Post, and more without warning. Hardly two weeks later, the quota management system in Google’s own cloud service platform (GCP) also failed, preventing users from logging in to their Gmail accounts and disrupting companies relying on GCP for their internal day-to-day operations.
Now, imagine if either of these outages affected core cloud functions of the Pentagon’s network or a 911 system. What do you think would happen? The country could be left defenseless. People could die. It would put the government far beyond the point of needing a wake-up call to improve the resiliency of their cloud infrastructure. In fact, that wake-up call has been happening for years now, with each new cloud outage leaving a more detrimental impact than those preceding it. Additionally, an increasing reliance on any single one of the ‘Big 3’ cloud providers on the part of government agencies or organizations migrating their services and/or data to the cloud could serve to exacerbate critical service failures in the event of future cloud outages.
One of the most prominent ways future cloud outages could be mitigated is for government agencies to disperse their cloud computing infrastructure across multiple cloud providers. This just means avoiding any cloud proprietary services, using open applications and services that will work simultaneously on multiple clouds. This could be done through implementing common cloud architecture in those providers’ infrastructure, which would alleviate the worry of a government agency or organization’s critical services being unilaterally affected at once.
The Problem With Single Points of Failure in the Cloud
Should cloud providers implement a common cloud architecture in their cloud infrastructure, such a move would allow interoperability across their infrastructure’s network. Subsequently, this would allow government agencies and organizations to distribute their cloud services across that distributed network, potentially (yet severely) mitigating the interruption to critical government services on the cloud.
The root problem of cloud outages disrupting critical services can only be solved by using more than one cloud provider. We have been creating high-availability systems for decades, and they all have the same thing in common: no single point of failure. Essentially, relying upon any single cloud provider equates to a single point of failure when a disruption or outage occurs.
However, this raises another issue regarding the pushback of cloud providers pitching their proprietary cloud services. These are high-margin offerings that keep their customers locked into their platform. Cloud providers even encourage proprietary technology in their certifications and training programs, so establishing a network of infrastructure that does away with the concept of a single point of failure is more a matter of education of the cloud architects, cloud engineers, and others who work on cloud platforms.
These workers need to be taught the benefits of designing multi-cloud services versus single cloud ones with proprietary vendor offerings, so it will only be a matter of choosing the interoperable services on multiple cloud providers. This is not only totally feasible but also in the best interest of the customer. This way, the failure of a single cloud provider will have minimal impact on critical government services.
Cloud providers already have the best technology and people in the world expertly designing and maintaining their systems. The problem is that, eventually and inevitably, all technology breaks. We must plan for what happens when this occurs. Does another cloud take over leave it to “business as usual,” or are we in a position where we will remain lost if a single provider has an outage? In order to address the root issues with cloud outages disrupting critical government services, the best and easiest solution is to disperse those services and the infrastructure supporting them across multiple interoperable cloud providers.
The author, Michael Gibbs, is CEO and founder, at Go Cloud Careers.