“Cloud repatriation,” like “cloud migration” and “cloud native,” is a tech term borrowed from the language of social science: all of these terms describe a relationship to a place of origin. What each really describes, though, is where someone, or something, lives. In social science, that someone is a person, someone born a citizen of one country or returned there after displacement by conflict or other political circumstances. In tech, the something born in or returned to its place of origin is an asset or a resource an organization controls: it’s your organization’s data, its software, or whatever else you need to store to be able to run it.
After years of cloud migration dominating the conversation about software and data hosting and storage, the term “cloud repatriation” is emerging as a new hypothetical for migrated and cloud native organizations. So many organizations are now hosted on the cloud that a greater number than ever have the option, feasible or not, to move off. Whether any cloud-native or recently migrated organization would actually want to move its resources back on-premises, to a data center, is another question. To discuss this question and its implications for the future of the cloud as a business solution, C2C recently convened a panel of representatives from three major cloud-hosted companies: Nick Tornow of Twitter, Keyur Govande of Etsy, and Rich Hoyer and Miles Ward of SADA.
The conversation was charged from the beginning, and only grew more lively throughout. Sensing the energy around this issue, Ward, who hosted the event, started things off with some grounding exercises. First, he asked each host to define a relevant term. Tornow defined repatriation as “returning to your own data centers...or moving away from the public cloud more generally,” Govande defined TCO as “the purchase price of an asset and the cost of operating it,” and Hoyer defined OPEX and CAPEX as, respectively, real-time day-to-day expenses and up-front long-term expenses. Ward then stirred things up by asking the guests to pose some reasons why an organization might want to repatriate. After these level-setting exercises, the guests dove into the business implications of repatriation.
The question of cost came up almost immediately, redirecting the discussion to the relationship between decisions around workloads and overall business goals:
Govande’s comments about “problems that are critical to your business” particularly resonated with the others on the call. Govande briefly elaborated on these comments via email after the event. “In the context of repatriation, especially for a product company, it is very important to think through the ramifications of doing the heavy infrastructural lift yourself,” he said. “In my opinion, for most product companies, the answer would be to ‘keep moving up the stack,’ i.e. to be laser focused on your own customers' needs and demands, by leveraging the public cloud infrastructure.”
These sentiments resurfaced later in the discussion, when the group took up the problem of weighing costs against potential opportunities for growth:
The more the group explored these emerging themes of workload, cost, and scale, the more the guests offered insights based on their firsthand experiences as executives at major tech companies. Tornow used an anecdote about launching the game Farmville at Zynga to illustrate the unique challenges of launching products on the cloud:
During the audience Q&A, a question about TCO analysis gave Hoyer the chance to go long on his relevant experiences at SADA:
As soon as the conversation began to wind down, Ward put the guests on the spot again, to ask Tornow and Govande point-blank whether either of them would consider repatriation an option for their company that very day. Unsurprisingly, neither said they would:
By the time Ward handed the microphone back to Dale Rossi of Google Cloud, who introduced and concluded the event, the conversation had lasted well over an hour, leaving very few angles on the subject of repatriation unexamined. Many hosts might have felt satisfied letting an event come to an end at this point, but not Ward. To leave the guests, and the audience, with a sense of urgency and resolve, he treated everyone on the call to a rendition of “Reveille,” the traditional military call to arms, arranged exclusively for this group for solo Tuba:
Repatriation may not be a realistic option for many if not most businesses, but discussing the possibility hypothetically illuminates the considerations these same businesses will have to confront as they approach cloud strategy and workload balance. “Nobody on our panel had heard of anyone born in the cloud ever going ‘back’ to the data center,” Ward said in an email reflecting on the event. “Any infrastructure cost analysis is a ‘complex calculus,’ and there's no easy button.” For Ward, there is one way to make this complex calculus manageable: “To get maximum value from cloud, focus in on the differentiated managed services that allow you to refocus staff time on innovation.”
When you hear the word “repatriation,” what comes to mind for you? What does it imply for your organization and the workloads your organization manages? Are there any relevant considerations you consider crucial that you want to talk through in more depth? Join the C2C Community and start the conversation!