 Hi Nikhil. This is Krish Raghuram here. Nice to meet up with you and thanks for taking this interview today. Thanks Chris. Okay and as you know I'm representing the product work group. We're doing the Newton design series because I work for Intel and I'm a planning manager engaged with Rackspace on the OpenStack Innovation Center. So why don't we start off by telling you, you know telling us a little bit about yourself. Absolutely. So hello everyone. This is Nikhil Kumar. I am the project technical lead for OpenStack Lance for Newton release and I work as a software developer for IPM and I you know engage in open source and open communities that sort of help OpenStack. So I'm focused around OpenStack and I try to make you know cloud life better. Now that is a good summary and as the glance PTL why don't you tell us a little bit about Glance? Sure. So OpenStack Glance as you may know. It's a reference implementation for the OpenStack REST API. It's a core service as defined by you know the branding that OpenStack gives and the requirements therefore. So that's what the major focus of Glance is like to make OpenStack infrastructure a little bit smooth for operators and users and then that's where developers are more than mostly. And I guess you know it is it's an API service so it provides feeds up, serves up images to Glance to Nova among other things and I think you also may have mentioned anything that was talking during the summit about Glance extending to also be a repository for templates and so on to be served up to heat. So is that where Glance is going? Absolutely. So Glance being a service that's being consumed by a variety of services in OpenStack. So we store basically images if you may but those images could be bare metal images. It could be you know production machine images as well and then volumes are stored as well. So over on Glance is a cataloging service if you may but it provides a couple and a data. So a couple of data and metadata on that data and it provides some level of coherency for that metadata. So we have some constructs that define what the data is and then the metadata is sort of a definition of that data and then it adds some extra information like a licensing or something like that. So that's where the core of the usefulness of Glance is and we are the mission statement of Glance is you know to provide you know upload and discover mechanisms so that you know data assets like heat templates or you know Murano catalogs can also be used besides what we already have like images and volumes. So that's where we are going. We also have some something called metadata definitions which is more of a textual representation of the metadata and it sort of helps the end user understand what that meta is about. So it's you can think about Glance as an information center for OpenStack. So it's like the first point where you go if you are in a new place. It's probably one of the main interfaces for the users in terms of uploading and using images and launching them. So we had a pretty successful design summit in Austin. So what were some of the hot topics for Glance over there? We have been focusing a lot on interoperability issues over the last year and that sort of continued during the summit design summit as well. We talked about more discoverable interoperable you know newly defined upload process that we are calling as import because we can add some level of features on top of upload to make it more operator and end user friendly. So that's what the focus for you know Newton is as well and we want to focus on so that's a veto only API and we want to focus on that to basically help users and operators in a inevitably longer term. So we want to duplicate version one of the Glance API and one of the dependencies for that was the adoption of Nova's version two Glance. So you know the Glance team was working very hard and getting that done and during the Newton you know as far as you know testing aspects. Besides that we are focusing again the end user experience is a little bit better and then you know evolving Glance is a unified catalog as well so that's being done in the glare repository or Glance repository but as a glare codebase. Yeah okay so that's that's great it's good to see the things that Glance is doing to move forward what were some of the user needs or main problems because we talked about features but what were the user needs or problems that you're seeking to address through this. So as of today's date end users have you know variety of ways where they can upload and store image data and metadata to Glance particularly image data because we have some advanced APIs that are operator defined but we are trying to come up with better stories around that and we are trying to come up with different mechanisms like asynchronous and synchronous mechanism to upload data into Glance in a unified sort of a discoverable more interoperable way so and that can be used by something like an open stack client or any new client that you're trying to build use that discoverability aspect of the API and then you know have a seamless experience for the end users if you may. So the work is focused around end users but from Glance aside we are building the you know building blocks for helping the clients to write those things. And we also saw that so we also saw some of the image sharing aspects are a little bit you know immature if I may you know you can do a few things with image sharing but the real vision of image sharing when it was really introduced was to provide sort of a marketplace platform so that you know dashboard can wear something more feature friendly for the cloud large scale particularly and for that to happen and to you know help the operators with the more more better contracts or more durability or manageability of the images that they provide we are coming up with the you know the next step for us is coming up with the concept of community share community image and then that will be like improvised image sharing you know sorry improved image sharing and that's that's something of a focus for Newton as well and end users centric too. Also there have been some cross-project efforts for better documentation so that's that's again one of the things that the team is focusing on. So you did talk about quite a few things there which I'm sure would have got initiated in the in the Newton cycle but probably have a life going way beyond that so what would you say at the top three or four priorities for the Newton cycle itself right or what you'd like to make make sure it's done before the next summit. Absolutely so one of the things that we have been working really hard for a really long time is the important factor again. A lot of has been discussed and a few of the you know issues that were blocking us during the past couple of cycles have been resolved to an extent. So having like a base for that refactor is what our focus is we have an approved spec that we are getting from last cycle and we are you know coming over the discovery layer and then you know a base concept for the upload in a new way. So that's one feature that we're focusing on. I mentioned about better image sharing and then the community image sharing concept so that's another feature that we're focusing on and we are also focusing on deprecation of plans version one so that you know operators and the users can easily move to plans version two when they sort of start adopting the better features that the version two has to provide and it's more friendly for a public facing API. So that's where that's what our target is and we also want to you know have some some sort of stability in the API that we expose for the artifacts or the unified catalog. So these three or four figures that we are targeting. Terrific and as you can I think you touched upon it a little bit already about you know improving interoperability and so on but what would you say would the key themes be for Newton? Could you kind of for this benefit of the listeners could you kind of summarize what the key themes would be? A lot of this information gets fed into you know product working group roadmap and stuff so we generally when we talk within the developer team or when we are having an interaction with the operator team we are thinking in terms of like a high level goals for the project as well. So you know we are already dealing with a lot of interoperability issues that we are addressing over this cycle and then going forward. So you know it's going to better provide a better experience for end users and the other services like I mentioned about porting nobody use version 2 which has you know trust concepts so that long-lived snapshots can succeed without without failing. So these things are you know important for us and then we that's the interoperability being one of the you know important things for Newton case. Besides that we do focus on a little bit of manageability because we are a core service and then you know if operators and users want to have a better experience for you know a largely adopted service and manageability is usually an important factor for Glance and I think while we are doing all of this we want to ensure that you know in the long run our code base is stable and we have a way to iteratively make profits. So we do focus on the modularity and then make sure our design and architecture are on the right track and keep a check so that the you know if there are any workarounds or if there are any temporary additions to the port base we have a way to address those in the next cycle. So that's one of the key themes for Newton as well. Terrific I think you covered a lot of ground there. So thanks for that. Is there anything else you'd like to add to close out the interview something for the audience? Yeah surely so we are stressing a lot on the version 2 of the API and I see that we have some future requests from end users and then from other teams as well. So I just want the you know community to know that we are working really hard on that with the interoperability and the discoverability layer of that. So all the input that we get from end users or other teams is going to be fed into all those themes and then we are going to make sure to you know have a big hand for all the major stakeholders in designing the very factor or any other enhancements that we do. Besides we are working very hard on the the layer or the artifact unified cataloging concept and a lot of services have expressed interest. So you know we want to see that done and I hope to get a lot of good good constructive critical feedback from the teams and end users about you know what's the best way to do it and then how would they like to see that going forward. Okay thank you very much Nikhil for you know taking out some time to do this interview really appreciate that. I'm sure many of the users and end users will also appreciate some of your the points you've made here. Wishing you all the best with the Newton cycle. I'm sure that the mid cycle is coming up and then the final release in October. I'm sure a lot of people have a lot of good things to look forward to from glance. So thank you very much. Thanks Chris I appreciate the interview and then I appreciate your time. Thank you.