Application portability with ROBIN Hyper-Converged Kubernetes Platform for Big Data
There is a level of portability. Rajeev Madhavan, CEO of big data specialist Robin Systems, said there is a fair amount of application portability between cloud systems with containers. “But unless you get data portability, vendor lock-in is inevitable.” Read more
Cloud Computing Means Vendor Lock-In, but It’s Not All Bad
Moving from expensive on-premise infrastructure to cloud-based systems is supposed to offer enterprises a lot more flexibility, but there are tradeoffs.
MOUNTAIN VIEW, Calif.—The allure of cloud systems, such as Amazon Web Services (AWS) and Microsoft Azure, is unmistakable. Companies can stop investing in expensive, hard-to-maintain data centers and move to the cloud where they only pay for the resources (compute power and storage) they need, and the vendor maintains the infrastructure.
The value proposition is unmistakable and a growing number of cloud vendors are cashing in. But panelists here July 9 at the wrap-up of the two-day AlwaysOn Silicon Valley Innovation Summit said customers shouldn’t have any illusions that by moving online to the cloud they are freer to move between competitive offerings—from AWS to Azure, for example.
“From hardware to just about the application layer, I would submit the cloud introduces a level of vendor lock-in we’ve never seen before,” said Abe Kleinfeld, CEO of GridGain, which specializes in in-memory data fabric. “Before you could choose the hardware, the network layer and more, and now it’s all abstracted. Is that bad? That’s up for discussion.”
There is a level of portability. Rajeev Madhavan, CEO of big data specialist Robin Systems, said there is a fair amount of application portability between cloud systems with containers. “But unless you get data portability, vendor lock-in is inevitable.”
The reason, Madhavan said, is that it’s not in cloud vendors’ interest to make it easy for customers to move from one cloud system to another. “If you put 60 petabytes of data into a cloud system, you are locked in. It’s not going to be easy to move, and it will be expensive. It’s a great opportunity for a great solution. CIOs want this,” he added, noting that his own company is focused on this area.
In the case of AWS, James Liao, CEO of Pica8, a software-defined networking vendor, said he thinks the company will develop better ways for companies to move their data off AWS when they want to, but it won’t be a total solution. “I think they will supply 90 percent capability to move your data; the rest will be vendor lock-in,” he said. “But there’s going to be interoperability.”
Vendor lock-in is less about systems that restrict portability, and more about the cost of switching, Liao said. “A vendor can’t prevent you from switching; it’s just a matter of cost,” he said.
To what extent and at what cost interoperability will come remains to be seen. Gaurav Tewari, managing director of Citigroup Ventures, said it’s not a question of yes or no. “It’s going to depend on the type of company.” He also said vendor lock-in isn’t a bad thing if the company is giving the customer what it wants. For example, he noted that security is a big issue, and being able to show the ability to recover from a breach or other disaster—natural or otherwise—is a very big deal. “AWS has a great story there,” said Tewari.
He also talked up cloud services as a great way for companies to experiment without making a large infrastructure investment.
Kleinfeld of GridGain said the question of interoperability gets confused by different interpretations or references to what constitutes a cloud. “Everyone defines cloud differently, whether it’s AWS, or the Google Compute Engine. We hear about the ‘open cloud,’ but I don’t think it’s here yet,” he said.