Welcome to the Episode 315, part of the continuing series called “Behind the Scenes of the NetApp Tech ONTAP Podcast.”
This week, Esteban Rubens of NetApp (https://www.linkedin.com/in/erubens/) and Jason Klotzer of Google (jklotzer@google.com) join us to discuss enterprise imaging in healthcare and how NetApp and Google Cloud team up to offer some of the best solutions in the industry.
For more information:
Tech ONTAP Community
We also now have a presence on the NetApp Communities page. You can subscribe there to get emails when we have new episodes.
Finding the Podcast
You can find this week’s episode here:
You can also find the Tech ONTAP Podcast on:
I also recently got asked how to leverage RSS for the podcast. You can do that here:
http://feeds.soundcloud.com/users/soundcloud:users:164421460/sounds.rss
Transcription
The following transcript was generated using Google’s voice to text transcription service. As it is AI generated, YMMV.
00:55 – 01:09
0
0.77
Hey Justin, thank you. I’m part of the health care team at NetApp. I cover cloud and AI Healthcare which process providers, payers and life sciences and you can reach me on email, Esteban Rubens or LinkedIn.
01:10 – 01:53
0
0.77
All right. Excellent. Also with us today from Google Jason. Closter is here. So Jason, what do you do and how to reach you, everybody? So yeah, name is Jason Klotzer. I work for Google Cloud specifically. I’m in the health care and Life Sciences vertical within Google cloud and even more specifically I work on Imaging related projects. So anything within the medical imaging space downstairs providers etcetera for my med tech that that is what I do for Google cloud and trying to basically come up with solutions that work. Well for, for our customers. You can reach me at Jay Klotzer@google.com, or you can reach me on LinkedIn as well.
01:54 – 02:54
0
0.74
All right, excellent. And as we hinted, we’ll be talking about health care and imaging as well as Google Cloud but to start off with, I just want to kind of level set and talk about Google Cloud itself. So, Jason off, you know, it’s pretty self-explanatory that Google cloud is a cloud instance that Google has. But can you give us a little more color into what Google Cloud can do and and how people can consume it. Sure Justin. So often in Google cloud is one of your massive hyper scalars as many people hear about it, but I think that one of the main things that sets us apart is the indirect approach that we have within the services that we offer. I mean, actually Google has been known for, you know, Open Source Products that have come out like tensorflow and could use that way kind of, you know, put out into the ethers. So others can kind of learn from from what we’ve learned, but I think the industry approach that’s really common about in becoming very wage.
02:54 – 03:54
0
0.77
Popular, nowadays is where we differentiate ourselves weather be in the finance District or we’re we’re talking about today Healthcare having services that really standardized within those jobs means is really, really what sets us apart. So TL, DR. Is we try to approach customers with a managed offering where they don’t have to worry about their own scale. They don’t have to worry about the availability of data and these sort of traditional on Prime concerns and we kind of encapsulated into domain Centric services that they can leverage as as close as possible to the way they leverage right now. That’s what we do. Yeah, and like, any cloud provider, you’re offering things like compute and storage and services as a service, right? So applications as a service plaza has a service and you’re managing it on your end. So that end users don’t have to worry about it. They can just consume the resources and then, you know, basically rent them out and put them back. Yep. That’s exactly right. So, often.
03:54 – 04:54
0
0.78
With that said, I mean, NetApp has a definite presence within multiple Cloud providers as well as Google. So Esteban, can you kind of give us a little insight into that relationship with NetApp and Google Cloud? Absolutely, our main collaboration revolves around data storage management, obviously, that’s what setup is well-known for. And the main idea here is taking on Tab app, which is the product that is at the core of a lot of what network does and making it available as a service in the Google Cloud. So the same on that that people are used to on-prem for data storage and management is available in the Google cloud. And the idea is that it’s kind of the embodiment of the data fabric concept of being able to seamlessly move data from Edge to court to Cloud within the same environment. Then again, the key word there is seamless. So we want to add a layer cake.
04:55 – 05:54
0
0.71
A huge amount of offerings that are within Google cloud with the stuff that we add value to write, which is basically file storage, and block storage. Particularly as we talk about Enterprise Imaging in healthcare. It’s file storage because that’s a big part of Imaging. You mentioned Imaging and Healthcare. Can you get go into a little more detail about what exactly they, you’re talking about with imaging? Cuz I mean, Imaging can cover a lot of different things. What does it mean in this context? Yeah, Jason and I both have history in in what way you can call Medical Imaging, or Enterprise images of Jason, please jump in whenever you want. The idea here is, would used to be in the domain of radiology, and Cardiology has really changed exploded throughout Healthcare. It’s very hard to find a domain within Healthcare specialty that doesn’t use Imaging in some way we’ve gone from birth.
05:54 – 06:54
0
0.72
Based Imaging like your your standard x-rays you for chest or you know extremities or what-have-you to demography. Like cats can mammography ultrasound memorized tried using basically a magnetic field to knock Adams off their access and with the radio signals instruct an image, all of that still holds true. Additionally, we have what some people like to call visible light Imaging. It’s like the standard photography, and video that off. Now, you see, if you get an endoscopy. There’s going to be video and images. If you go to the dermatologist, and you take a picture of a mall or some other feature that they want to study in our thermology. Pathology right. Is is now the the big hot area for Imaging. That is not really digitized yet and it’s starting to age.
06:54 – 07:54
0
0.75
Go in that direction. Moving from, just having glass slides, do digitizing, those slides and making those pathology images available for collaboration analysis. May be running down through some algorithms, mean, the list goes on and on and on. So, that’s why people now talk about Enterprise Imaging, within Healthcare to denote that, it’s really the collection of all Imaging. And the idea of how do you manage all that data? It’s a lot of data, some, some estimates are between eighty and ninety percent of all Health Care data is Imaging. So it’s a lot of data that has to be stored. Managed secured has to be made available to the people who use it to do. Patient care to interpret the images. Dictate reports. It is a really huge booming field. Globally. It’s not just in the US and and Europe it’s everywhere. I’m not going to go through the gamut of Mojo.
07:54 – 08:54
0
0.77
I used to that extent. Things will probably be here for a long time. But I, I think some of the points I want to make about Enterprise Imaging, similar to to what Esteban was just saying was, how it’s really taking off right now, you know, Enterprise Imaging. You just kind of derived from the two words. I mean, it’s Imaging within the Enterprise domain. So often think about it as a hospital or multiple hospitals, more likely with multiple outpatient imaging centers, multiple different ways in which aging is acquired within each one of these and all the work flow around, you know, the acquisition, the interpretation and even Downstream functions of the Imaging data off. So that means that you have a lot of network, a lot of storage, a lot of compute, and a lot of business applications, kind of interfacing, with one another and a lot of actors, you know, dead.
08:54 – 09:54
0
0.78
Actually working with that data that that I would say, is kind of the encapsulation of Enterprise Imaging, and I also, you know, think just giving that sort of a description on how fast it is right now. That’s really where these new sort of architectural tools, which is what I kind of consider Cloud to be quite honestly, that’s where it becomes a really long, interesting. And one other one other just comment here on the storage side. Naturally. A lot of these Imaging modalities are Imaging devices becoming better and better as well as the bomb mentioned Thomas, emphasis for instance, in the mammography space inherently because they’re getting so much better off. The data size itself is also becoming much larger and the resources that you require to process this sort of data in a similar time frame to what a, you know, real.
09:54 – 10:54
0
0.76
Ologist or mammographer, more specifically would expect those become more intensive as well. Look at CT or MRI, you know, they keep on pushing out more and more slices with a higher resolution digital. I mean similar to what Esteban was saying, you’ve got opthamology, you know, Dermatology digital pathology, you know, even Factory all of these spaces are looking at was what has happened? Classically and Enterprise Imaging, and they’re kind of saying, you know, we want to do that as well. We want to be able to have that sort of a capability to share this information, you know, across physical locations, to be able to do analysis with this, this data and it all makes sense quite honestly, so I think there are a lot of operation within that domain and more crisply. It also points, the fact that this massive amount of data really needs to have standardization around it and and birth
10:54 – 11:54
0
0.74
That’s where a lot of the kind of things that we’re going to talk about coming to play. Yeah, and one more comment in terms of cloud. Think about most health care organizations in the majority of the world have been digital for, maybe about 20 years, give or take. There’s a huge amount of archived data that used to be considered maybe a liability. And and people used to talk about while we are going to be able to delete some of these images, as soon as maybe regulations allow us and in the last few years. I think that’s drastically changed because there’s more and more recognition that there’s a lot of value in those images for patient care, for clinicians home. That can only be unlocked if you do further analysis, so, this huge mass of historical images that used to be perceived as not very long.
11:54 – 12:54
0
0.89
Call is now gold, mine. So the the question is, how do you use it properly? How do you alleviate some of that footprint that exists on Prem? And people maybe don’t want to have fun Prem while still keeping the images available and in a place in which that analysis can take place in a in a reasonable way in terms of performance. And if you think about what cloud does that ability to elastically consume huge amounts of resources, whether it’s computer storage, that’s exactly the use case. So, that’s why there are such a good fit between Palm Imaging, where it’s at today in 2022 and were were claudi’s. Yeah. I couldn’t agree more. It’s, it’s interesting. The state of image as it is now and you look ten fifteen years ago at precisely to your point Esteban, most Imaging organizations back then were talking about what they would refer to as in dog.
12:54 – 13:54
0
0.74
Lifecycle management or or more specifically how to Archive out Imaging data, that’s no longer necessary for a diagnosis to be able to save on, you know, on Prom costs. But now I I’ll be honest, a lot of those who are contemplating, that sort of a choice, Mario, probably haven’t implemented, anything maybe because retention rules, were so complicated for them that now they’re actually happy. They haven’t implemented anything for a because now they can actually leverage that massive amounts of data for improving their workflows quite honestly, which is where it comes into play. So, yeah, I one hundred percent agree with that perspective. Yes. Sizes of images of really exploded over the course of the last, you know, decade or twenty years. I mean, we’re looking at 100K images all the way up to like. Now ten gig twenty gig image has right? Like they’re, they’re massive wage.
13:54 – 14:53
0
0.81
The megapixels are much higher. The resolution is much better. So the other issue is that that data kind of stays in place for a while. Maybe it doesn’t get access for a while. So now you’ve got all this expensive or storage on site and you’re powering it and your cooling it and you’re not really using it. So you got to figure out a cheaper, way to handle that and that’s also where the cloud comes in as well as is the archived. Here’s right. And you know, think about pathology if you think about the very standard pathology slides that that’s digitized, we’re talking about gigapixels. It’s an inn. This is just one slide, whereas in a case, you may have a multiple sides and you’re going to have thousands and thousands of cases and certainly in the larger institutions. It’s tens of thousands or or more. So the amount of storage that he needs really gross very quickly and on top of that another maybe complicating factor is that these images are very compressed, right? Whether it’s jpeg jpeg2000 wage.
14:54 – 15:53
0
0.72
Angie. There’s some format that is going to introduce some kind of either compression or if it’s not compression. It’s going to be some kind of a, the most you can do right to improve the storage of us images. So, the standard approach are saying, well, we’re going to just apply more compression later on, doesn’t quite apply wage. Scale is really important, right scale and in the ability to move data very quickly and easily because sometimes you need that data back, so I don’t want to have that data offline or even near line in the sense that it could take a while because as Enterprise Imaging progresses, I’m sure this is going to be addressed. But in the majority of cases, the applications are still very sensitive to latency and they don’t have a clear idea that maybe if you want to really take advantage of something that’s much cheaper, it may take a minute or 30 seconds.
15:54 – 16:54
0
0.72
Five minutes or whatever. The case may be, you’re going to get a timeout. So it’s very important on the application side that, you know, if you need to pull an image, that’s ten years old for whatever reason, right? It could be because there is a legitimate patient care need or there’s a medical legal requirement or there’s a research requirement. We don’t really know. You have to be able to say I woke image and get it back without going into a complicated workflow. That is based on interruptions and it’s going to throw errors in the application. Guess you’re not going to need a new image are going to need millions of images. So it’s very important to be able to do this very seamless scaling and then also choose whether the compute that are going to use V to display process or do anything with those images is going to be in your data center on prime or in the cloud. So it’s really about this hybrid IDE.
16:54 – 17:02
0
0.88
Then it’s going to be spanning the customer’s premises and the cloud, and that’s what makes the most sense for most people.
17:04 – 18:04
0
0.77
Yeah, so they’re kind of architecture as to as to balance describing right. Now. This hybrid architecture is, what I would say is, is most popular within the, the prize Imaging domain right now, especially for customers who are looking to leverage what cloud resources I would say that purely cloud-based diagnostic tools are not the norm. Yet that it absolutely is more along the lines of having traditional clinical, applications interfacing, with a cash on-prem for fast access for, for access to data. That’s more in proximity to the diagnostic sort of, you know, scenario and that all of the data that may not be necessary. For maybe the patients that are being interpreted that day or being seen that day. That
18:04 – 19:04
0
0.81
Data is likely going to be pushed over to more elastic, you know, storage and compute environments like a cloud provider. That’s that’s a very, very typical hyperbole architecture nowadays that that we’re seeing, which actually has multiple benefits, right? I mean, if you look at it from a secondary use standpoint, I, you likely don’t want to run your ml, you know, training or maybe even your your inference inference and prediction workloads, you know, on-prem you may not want to have to deploy all of that hardware on-prem and maintain it, which is why a lot of those workloads. If you have a majority of your cloud in in hyperscale or like in a cloud, for instance running those workloads, you know, around ml is is more natural. I would say. So it kind of fits well together from from an architectural standpoint wage.
19:05 – 20:04
0
0.80
Yeah, Yeah, from the point of view of not Reinventing the wheel, right? Because a lot of people find it kind of overwhelming you maybe know the data science but then you have to build all these Investments and maybe compile or configure some open-source stuff for as having access to those environments. Ready? Ready made. That’s what people want because they want to get to whatever problem, they’re trying to do a couple and not have to figure out how they get started And maybe invest months, right, or weeks or months to get to the point of being able to start. So, that’s certainly one of the values that we see in a relationship with Google Cloud. But they already have all of this set up, right? So you go to your counsel, you just spin it up and there it is. And then you go to choose your data set and you can start using your data and getting insights, I I’ve seen some of the demos and they’re actually pretty cool in that regard, yeah, so the, the service, I think you’re referring to us to bond as the healthcare API, where we offer.
20:04 – 21:03
0
0.71
Add item store h7v to store and a fire store. And as I was mentioning earlier, you know, focused around standardization. This is really where I think, you know, we, we kind of, you know, put put her opinion on the table. If you will. That we absolutely believe that standardization around these Services is, is a direction. It definitely Foster’s interoperability between business applications. And it also allows us to more easily integrate with, in fact domain, you know, we don’t need customers to implement some proprietary anything. It just allows them to pick it up and start using it, which is really phenomenal. Quite honestly. I’m just bringing back to the thought of just Enterprise Imaging as a whole and how this sort of hybrid architecture fits in home.
21:04 – 22:03
0
0.77
Specifically, within the Enterprise Imaging domain, larger sort of partakers within that domain with typically break up their systems in a way where you might have a packs, which is a picture archiving Communication System. Basically the closest point to clinical care when it comes to diagnostic interpretation off, and then you would have a vendor-neutral archive, which is really your application point, for all of your Imaging data across your Enterprise Imaging domain. The reason I’m kind of explaining this is because that sort of a Enterprise Imaging architectural pattern fits, very well when you lay it on top of what’s available in a hybrid architecture, nowadays between your on-prem sort of Fast Access, you know, cash kind of a hardware and your name.
22:04 – 22:42
0
0.75
A scalar availability within like a Google cloud provider as an architectural pattern. So basically on-prem I would say that the equivalent would be your multi-pack systems for your immediate clinical care and then having the aggregation be a VNA, which could likely live in a cloud provider. Since that’s not where your, your clinicians are directly accessing. Get in most cases and they can also Foster secondary usage, whether it be machine learning analytics, and things like that. I think the, the architectural package very well to, to what we see now.
22:44 – 23:42
0
0.75
What about other use cases of cloud and Enterprise Imaging that are more mundane? Maybe but no less important like image-sharing, right? And they need protection from. We see now an explosion of attacks on Healthcare institutions. You read about it, everyday, ransomware, and other attacks. It’s so the idea that you can have copy of your images safe and you can keep working or go back to work, even if you are targeted and then yeah, maybe image-sharing in the sense of well, maybe I was patient and I think anybody who’s been a patient in the last few years and has had any Imaging, you know, they can be a pain to get your images and maybe you get a DVD or some kind of a physical piece of media that can get misplaced and then you get a bunch of papers. Just we’re still in the Stone Age when it comes to that and this is something that certainly the cloud was born for.
23:43 – 24:37
0
0.75
All right, one hundred percent agree. Yeah, I mean the mundane the mundane use cases I think, are really what? Initially drive a lot of this, especially from an adoption standpoint. So just, I mean, the simple use case of backing up Imaging data to some place. You look at the storage offerings right now from found objects storage standpoint and you using, you know, a configuration such as an archive store. I mean, you can store GB terabytes worth of data for, you know, PV to a dollar. So I absolutely think they’re popular, maybe still under leveraged. I would say wage. I think when one of the one of the problems maybe in that domain is
24:38 – 25:38
0
0.72
Maybe folks try to do too much at one time. And then this is completely my opinion. But mundane use cases are are really things off that exist. I mean, I would say the Monday news clips that were describing right now, are things that can be enabled very quickly right now, in in the services that are available. Like, I mean, even the sort of work that we’ve done together, you know, between that up on top and and Google Cloud, I mean being able to just simply back up data to to the cloud. I mean, it’s it’s a configuration tasks at this point, you know, those kind of things are extremely impactful. So I guess what I’m trying to say from a tlg, our standpoint is dead, maybe customer recommendation is take a phased approach where you can enable use cases one phase at a time instead of a big bang approach off.
25:38 – 25:53
0
0.84
And I think that could, absolutely, you know, bring folks and allow them to achieve their their goals much faster than trying to achieve everything. At once. It’s the way I would kind of stayed it.
25:55 – 26:55
0
0.70
Yeah, Justin, I’ll send you a link that we can post married to this safer. Graphic. We have that talks about the cloud journey in healthcare, especially in Enterprise Imaging as it’s not a linear path. There’s all these things, you know, Jason’s point that you can start anywhere. There’s just so many ways to get into it and it’s not something that you want to do for the sake of technology, or because somebody told you that you have to go to Cloud. But there’s something that makes sense for the patient care aspect for clinicians to reduce mundane tasks and off, increase the, the ease of sharing data among people who have the right to see it. So there are many, many benefits, a lot of organizations, a, maybe a little overwhelmed. It’s like, oh, you know, it’s a big problem. What do you do? If you break it down as usual, right? To manageable chunks this idea of of a journey that you can start off.
26:55 – 27:26
0
0.79
At Point ABCD doesn’t really matter and then you can pick up where you left off later on makes a lot of sense so we can we can share that as well. Yeah, another use case that we haven’t really covered is you know, the the locality of data rates. So a lot of these doctors offices, May report back to a centralized office where all the data is being stored and they need to access this data, but they don’t want to keep it off. I’m necessarily, right? So that’s that’s another place where the cloud fits in because they can access it without having to set up a whole infrastructure to do that.
27:28 – 28:26
0
0.77
Yeah, that’s that’s absolutely right. And they’re they’re I would I would say as there are a couple of different levels and types of access phone, not from like an authorization standpoint. I mean that’s you know implicitly there. But the types of storage that they want to access, who the actor is, you know, a proximity of that information to them. There are a lot of details that, you know, we can go into there but, you know, Justin’s your point. I would say that there’s the location of the data from a sort of data center standpoint, which naturally I think the cloud lends itself very well to that because there are usually very strategic points where these data session success where customers can basically just configure, you know, where they want their data to be replicated and when and how, and then there is the, the on-prem cash portion off.
28:26 – 29:26
0
0.77
Which, you know, as we mentioned a couple of times exists in the health-care domain, you know, those different. Those different levels of storage can all be configured to make sure that the the information is available to, you know, every single type of actor within the health-care domain, whether they be the diagnostic actor like a physician, or to a, a principal investigator within, you know, the research domain for instance. They want to, you know, as Esteban was alluding to earlier run like a large batch workload on a massive amount of Imaging data wage, you know, these different characteristics of storage. Can all be configured based on the way in which folks wanting to work with it basically, so I think there are tremendous amount of options that we have nowadays, and I would, I would absolutely say that they, they fit the majority of the use cases that we’ve seen within the health-care domain those far especially around Enterprise images.
29:27 – 29:46
0
0.77
Especially with this hybrid, sort of a a nature that we see and also because of your Global footprint, even things like responding to regulations for if you’re in the U the home or, you know, you specifically to a country, the data can’t leave that country or any any sort of like right now. Exactly. Yeah.
29:48 – 30:47
0
0.75
Yep, you’re absolutely right. I mean, this, this happens all the time with various Healthcare Providers. I mean, it’s it’s almost like a script conversation that we have with them, which I’m sure you, you do as well. We’re off. We, you know, we we kind of determine. Okay. Well, what are your data data regulations, you know, can they leave the country? Do they have to be in a specific data center, you know, off the visibility of the data from an auditing standpoint, you know, these are all the kind of concerns that we’ve, you know, had to meet over and over again. And then you go to the specific bodies that have to make sure that you meet as well, you know, whether it be from a, an EU standpoint or M gdpr or, you know, it might be FDA, it might be HIPAA. I mean, if we talk about Federal Regulations, we can talk about fedramp. I mean, I would say another benefit within this domain and going with an enterprise-level service off.
30:48 – 31:47
0
0.73
Whether it be one that NetApp offers or one that Google Cloud offers or a combination of the two, our, our businesses have been around the block many, many times within this domain to know that these are all things that we have to keep, you know, Loctite and make sure that we are abiding by all of these different regulations, verse a home, you know, build your own type approach. So, I would say that that’s another one that comes up quite often when we talk to customers. I would imagine another thing that comes up pretty often, is, you know, I’ve met that I want to use cloud. How do I get there? Like, how do I move to the cloud? So, how is this? How are these EI data sets? Moving to the cloud these days? Yeah, that that’s a really interesting question because both ends up being a deterrent for four people to get to the Cloud. A lot of the more standard applications were built not with cloud and mind. So you need job.
31:47 – 32:47
0
0.73
Nectar, you need something that allows you to to get the data where needs to go. As Jason, was referring to, before in the packs world that the application that’s closed to the choice of providers Radiologists. Those are it’s starting to to be more mindful of the cloud that space. But, but by and large, it’s not quite dead and DNA’s are farther along for sure. So it’s important to have that glue to have the whatever connectors, middleware, whatever you want to call it. And that’s certainly where we came in. We make it easy, whether you’re a customer that has on pepper already and then we can use things like snapmirror, right? You can go from your mom from ontap to the cloud, to the instance on the Google Cloud, using the same technology that you are using before. So you don’t really need the money.
32:47 – 33:47
0
0.71
Location to quote, unquote, allow you to do that because you can just get the data there and if you don’t have on tap, you know, your poor soul, right? Who doesn’t have on tap on Prem, then we have SAS offerings like Cloud sync. That allow you to go from any source to any Target, really, and it’s all about simplifying that appointment of data. And once the data is up in the cloud, then you can do whatever you need to do with it, whatever you want to do with it, but it’s it all goes back to the idea of wage data Fabric and and easily and seamlessly moving data. No matter where it is. So we don’t really have a preference. It doesn’t really matter. And I think from both the net up perspective and I’m guessing it’s the Google perspective as well. We care about the data itself and that the data is useful in this protected and it’s doing what it’s supposed to be doing and then we build all the table.
33:47 – 33:57
0
0.71
Laying around that on all the features to allow that to happen again independently of the details. So we we can abstract a lot of that detail.
33:59 – 34:59
0
0.89
Yeah, so I 100% agree with the, the approach Esteban mentioned. I mean, considering it from a data standpoint, considering it from an application consumption standpoint. I think absolutely one of the areas where, you know, NetApp and and Google work together and and being able to leverage, you know, technical capabilities like on tap wage that really is a differentiator, you know, being able to move the storage from one place to another and not being forced into an application layer. That that’s really true. I would say that that is really a wonderful way to be able to move information when you talk to customers. I would say that you also need to discuss their thoughts of cases because I mean customers typically don’t jump at the opportunity for purely Technical Solutions. I think as Esteban was alluding to as well, you know, birth.
34:59 – 35:56
0
0.77
Talking to customers about, what are their immediate concerns, maybe flavoring in a little bit of futuristic things, that might be interesting to them. I think, you know, the one, we probably all think about Iraq ml or secondary usage of data, or, you know, the availability of analytics on their data, you know, being able to have a use case in my life. And I think, from from any sort of a solution architecture standpoint is is key. Right? When you have that use case, or those several use cases that are really enablers for a for the customer from maybe from a technical standpoint, but mainly for a business opportunity standpoint, and and from in this domain from maybe page Centric approach where you can say, okay. Well this is actually improving patient care from when you have those use cases. I think that’s
35:57 – 36:57
0
0.76
Where you want to be. Once you have those use cases then breaking down into the sort of technical solution and how we can enable this. You know, let’s say if we do just want to be able to replicate data. You have an existing on tap, you know, installation. Cool. You’re you’re able to replicate it to Google Cloud just, you know, just like that. I don’t know me better heard that, but I sat my fingers. So, you know, being able to being able to, you know, enable use cases by a Technical Solutions is really dead where I would position, you know, the packs and being a conversations as well. And then you’ll naturally see that there are many different variations of of services within a, you know, a combination of Google cloud and NetApp that can support the needs for these customers, whether it be purely, you know, a a storage play, whether it be a, you know, machine learning play a game.
36:57 – 37:57
0
0.77
Got a machine learning and analytics play, you know, there are many, many different variations, but I would say to start there and the solution will follow. Yeah, and it’s interesting to start with a pain Point, such as how to protect your data against ransomware. That’s something that every everybody involved in this conversation will be really concerned with whether it’s a really a g administrator whether it’s a c. I o c m a o of course the CCS everybody is just a really huge problem and it’s been a huge problem and will continue to be for the foreseeable future. So you take that you start with a pain Point. That’s very real something like that. And then maybe you add on something like what we want to reduce your thoughts, are our footprint on Prem that common refrain in the healthcare industry that I hear from customers. All the time is we want to get out of data Data Center business. It’s not our core competency wage.
37:57 – 38:56
0
0.76
Especially when you’re you start adding additional data centers for disaster recovery for business continuance. So there’s a big pain Point. There’s further pain point or an interest in getting out of the data boosts, our business and then this may be less pressing but still interesting idea of that secondary use of data to extract more clinical value, age that data and improve patient care and help clinicians. All of that can be combined by putting the state of the cloud. So that’s a very interesting way to frame a conversation because we know that that’s actually every Health Organization in the world is going to have these concerns. So you don’t have to sit down and ask them what they’re worried about. We already know and it’s much better. It certainly I talked to enough Healthcare cios who say the same thing. When we work with a partner. We don’t want to start from scratch. We don’t want to have them. Ask what your problems are a.m.
38:57 – 39:57
0
0.83
So, we have to explain it. We expect them to know already. And so this is a way to show that, you know, that these are problems that exist. And not only are you aware of what problems exist? We have a solution and better, yet. We have a solution that combines all these together. So it’s a pretty winning proposition. So one of the advantages of cloud is that it offers the ability for someone else, to manage storage in to set things up. And again, the platform-as-a-service concept. I know that Healthcare Providers are getting more and more interested in Ai and ML and their environments wage, but they might not have the expertise or the desire to pay for the expertise to set up. Those particular use cases, right? So, you know, set up their own ml learning, form. Does Google offer, like an a, i a m l as a service offering? Yeah, suggest an absolutely Google does. I think, when customers come to Google I would I would venture to guess.
39:57 – 40:57
0
0.77
And a lot of them come to Google because they’ve heard about Google in the middle space. So naturally within the medical imaging domain in Google Cloud. We have a whole gamut of different capabilities job offer. So, the, the sweet we, you know, the product name is vertex AI within Google cloud. And I can tell you that my job probably about 40% of, it is interfacing with customers who are interested in AI in this domain and the tools that we offer will satisfy the needs of everything from the researcher who wants to create their own pytorch or tensorflow, you know, machine learning, you know, sort of training approach, to those who really don’t know, very much about AIML at all. And may just have a set of images from like a CT study and want to see what they can log.
40:57 – 41:51
0
0.78
Get a machine to learn to automatically detect something, and that would, that’s what we refer to as automl. So yeah, my my suggestion is absolutely to look at the the future set within vertex, AI from a machine learning standpoint because this is absolutely something where we’re seeing on the Google Cloud side. A lot of Attraction from customers from every single day, you know, perspective and an attraction to using that for their medical imaging workflows, especially when it pertains to ml. So, yeah, a question. Yeah, it’s also good cuz they can invest in trying it out without having to invest in buying all the stuff, right? And if they like it, you know, yeah. Yeah, maybe they can invest in a later on and create their own environments, but for Google Cloud, I mean they have the expertise there to work with.
41:53 – 42:52
0
0.73
It’s a very, very good point. We do have customers who come to us and really quite honestly don’t even know to ask the right questions. You have your choice of your type of customer who comes with a use case in mind and they say, Hey, you know, it would be really awesome. If we could automatically detect so-and-so, or that we could automatically, you shall achieve this workflow based on this kind of data that that would be one group of customers who come to you, where you may even want to engage Professional Services where, you know, naturally Google has Professional Services who can build bottles for customers as well. If they really, are are absolutely hands off when it comes to that. But if they are someone hands on and are interested in learning starting with an auto and mail is very easy to do. You just plug data in allow automl to do its thing. And Bam, you have a dog?
42:52 – 43:52
0
0.74
Called you have your inference and then, as you get more and more into the platform, maybe you higher data, science, team, or some specific individuals who have a big science background. They themselves can Implement their own machine learning algorithms within vertex Ai and create pipelines where they basically are replacing the Autumn L automl with their own future, yielding, algorithms. And there you have it. You have some very specific AML workflow, that your team has created. And yes, all those capabilities are available in vertex, AI nicely integrated. I can say with Health Care API and bigquery. So Healthcare API on the side of NetApp would be the injection pipe off your Imaging data. So it all fits really nicely together. All right, excellent. Sounds like we got a lot to think about for Google Cloud as well as NetApp’s engagement with this space like the the Enterprise Imaging space. Yep.
43:52 – 44:05
0
0.79
So again as the bond, if we wanted to reach you, how do we do that? I’m at Esteban.Rubens@netapp.com or LinkedIn. All right, excellent and Jason.
44:07 – 44:50
0
0.74
Yeah, so Jason Klotzer or the first initial of my first name klotzer, my last name at Google or LinkedIn. All right, excellent. Thanks so much for joining us and talking to us all about Google cloud and Enterprise Imaging.
Why Is The Internet Broken? 
Pingback: A Year in Review: 2022 Highlights | Why Is The Internet Broken?