Flexible solutions for operational transformation -- such as multicloud, OpenStack's "modular" approach and edge computing -- are becoming increasingly popular among telcos as they work to meet the rigorous demands of 5G, says Ildiko Vancsa, an ecosystem technical lead at the OpenStack Foundation.
In a recent report the OpenStack Foundation reported that unique OpenStack
deployments had skyrocketed in the past year -- up 95%. On-prem private-cloud deployments and converged multi-cloud deployments have similarly continued to grow among OpenStack users, according to the report. Telco Transformation recently discussed these trends with Vancsa in the context of the telecom industry's preparation for 5G.
Previously, in part one of this two-part Q&A (lightly edited for length and clarity), Vancsa emphasized the flexibility demands of 5G while opining on their impact in the Asian markets and on cloud strategy. (See OpenStack's Vancsa on Adapting to the 5G & Converged Cloud Future.)
Here in part two, Vancsa talks about how the respective considerations of cloud strategy and 5G conjoin at the edge.
Telco Transformation: How do you see multicloud as a potential enabler for 5G? And/or vice versa?
Ildiko Vancsa: Multicloud adoption recently went up, and it is a really good concept. It helps us to give a bit more focus to our APIs. Multicloud is becoming something that we are trying to suggest to OpenStack users to consider -- to stand on multiple legs -- as opposed to putting everything on one setup and then failing if something goes wrong with that one.
TT: To speak of edge computing, 87% of respondents to one recent survey reported that they expect OpenStack to manage edge infrastructure going forward. In this context, what is the relationship between 5G and managing the network edge?
IV: Recently, we have been working on making OpenStack modular for those operators and providers who pick OpenStack for their environment. So you could pick up whichever service you need from all the projects that OpenStack offers today. And we are also putting more emphasis on making OpenStack itself as an integration engine -- so you can plug in different components, like storage back ends or SDN controllers -- to step one step closer to the telecom and NFV area.
So with 5G and edge computing, I think our goal is to be able to provide services that operators and service providers can use in their environments. First, so that they don't need to look into different projects and software tools to enable the different functionality, and, secondly, to address the different requirements of 5G and bringing edge computing into their system. They can look into one big open-source package and pick the components that fulfill their requirements, and put it in the system.
And as we see in surveys that are related to OpenStack in the telecom space, companies are seeing benefits for 5G. Also, they are using OpenStack as the base of IoT and edge computing to offer services more quickly while responding to constantly changing service demands quicker. They also like the reduced costs along with the reduced latency.
We see these components as the preferred and recognized benefits of using OpenStack. And from the Foundation perspective and the community perspective, we are putting more emphasis and effort into not just recognizing all of the requirements, but also to help the organizations to get involved and to ensure that all those requirements are addressed.
For edge computing, we have a new working group under the OpenStack Foundation umbrella -- so it's kind of on the same level as OpenStack. It's brand new. We had two calls and just spun up an Etherpad [document] and started to work together with the people in a more organized way.
This working group is focusing on gathering the use cases and identifying requirements. For those items on which the participants feel clear enough, we would like to move the momentum into the design and development phase as soon as possible, because this is a demand that we see from the participants -- both the individuals and the companies/organizations. As some of the companies are running workloads in production already, they would like to see more and more requirements addressed, both as related to OpenStack projects and as regarding the integration between OpenStack and other related technologies -- like, for example, containers.
TT: The instinct might be that public clouds would be more efficient for edge computing, but organizations seem to be moving more toward private clouds. What's your take on this? How are the respective trends of edge computing and on-prem private clouds reconciled?
IV: I think it depends on the use case. How I personally look at it, for me: Edge computing means the technology, and private versus public cloud is more of a deployment sort of thing -- although edge computing obviously has many deployment-related aspects. We get many different use cases where they would like to use edge computing -- and use cases vary from having something in a cell tower, to putting a device into shops, like a coffee shop or a bakery, or using edge computing in satellite communication areas. So I think the use cases will somewhat define the sort of cloud deployment, and whether organizations will use public or private-cloud deployment to fulfill the requirements of the use case. I would think that, probably, the landscape will change in the upcoming years as we see more production deployments.
As a summary, it depends on the use case and also the environment that gets utilized whether it will be private or public cloud. What we have seen in the OpenStack User Survey especially is that the on-premise type of private clouds are in high demand right now.
I think next year it will be either the opposite again, or just a similar discussion that while one is coming up the other is coming down because if you look back across the years, I think it's been a constant battle between public and private cloud. And who wins that battle?
— Joe Stanganelli, Contributing Writer, Telco Transformation