¶ Google Cloud Next Recap and AI
I don't remember which episode it was , but , tim , you were going off on Amazon . Q got announced right , because there were so many limitations . It's like it was super expensive . I think there was only . You can only ask you funny questions .
It was yeah , it had that whole yogi on the mountain thing where you can only ask it a few questions a day or something .
Welcome to the Cables to .
Clouds podcast . Cloud adoption is on the rise and many network infrastructure professionals are being asked to adopt a hybrid approach as individuals who have already started this journey . We would like to empower those professionals with the tools and the knowledge to bridge the gap .
Hello everyone and welcome back to the Cables to Clouds podcast . I am your host this week . My name is Alex Perkins . I am at Bumps in the Wire on X . I'm joined by Chris and Tim , my two lovely co-hosts . As always . Chris Miles is BGP main at BGP main on Twitter . He's not on X , he's on Twitter , Same with Tim McConaughey , who is at Juan Golbez .
They are both only on Twitter and I'm the one on . X . Thank you so much yeah .
I appreciate it .
Happy , very happy . Yeah , hopefully they ban X . They should send a bill to biden to ban it being called x .
I would support that . You don't say it might be happening , all right , uh , so we haven't done a roundtable in a while . Um , you know , besides , obviously , we do the news every two weeks and that's kind of a roundtable , but it's not about a specific topic that you know we can riff on .
So we thought this was a great time , right , google Cloud Next happened a couple of weeks ago . It was April 9th through 11th , so what we're ? It's like the end of April right now . So what ? Two , three weeks ago at this point , we , you know , we did an episode last year covering AWS reInvent .
You just like a recap of what we thought about a lot of the announcements . So this is going to be a similar format . Right , it's not going to be like a super long episode . We're just going to kind of go over some of the announcements that we heard from Google Cloud Next and give our thoughts on them .
It was also , I will say it was very AI and dev focused , right , big surprise there , surprise , surprise , shocking . I think I wanted to bring this up . It feels like , you know , the devs move the needle more than anyone , right , nobody cares if infrastructure people come in and they're like oh , I like doing infrastructure on AWS .
That doesn't matter , it's , whatever the devs , whatever tooling the devs like that's where the business is going to .
That's where the money is made . It's the revenue driver . That's it . The devs drive revenue and that's why they get what they want .
I mean , I get it , but at the same time I wonder if all this AI stuff is really hitting the mark for the dev people . Are you hearing these AI announcements You're like , oh yeah , that's good , that's good stuff . We AI announcements you're like , oh yeah , that's good , that's good stuff .
We got to get that , Because to me it's only been going on for , you know , several months at this point . But it's like elevator music to me , dude , I'm just like glaze over , but you know .
Yeah , I mean we don't care about some of the same things , right ? Like we're going to talk about the token limits on some of these models . That doesn't matter to us that much , right , but it probably matters a lot to devs . Yeah , but it's also a race that is going to just be constantly going back and forth . So how do you pick one ?
Feels like a race to the bottom sometimes .
Does Azure have a big event like this as well , like AWS and GCP Ignite , but it's not as well .
I mean , I guess this one it wasn't as cloud-focused either to be fair , but yeah . Ignite . So maybe we'll have an Ignite review at some point . We'll see .
Kind of depends what comes out of there , right ? If they'd only talk about developers at developers , developers , developers at Ignite , there won't be much to cover . True developers , developers , developers at Ignite there won't be much to cover , true , true , true , all right .
So , before we jump in , if anyone listens to this and they have feedback , something you think we missed , please reach out . Let us know if there's something you think we should have covered or we missed on or you wanted us to elaborate on more . Some of this could definitely be evolved into future episodes as well , so let's just jump right in .
So , uh , the first announcement I think I want to cover is this gemini 1.5 pro model um . The big announcement around this is really that it's yeah , pro , it's got to be the pro the . The big announcement is that this is one , one million tokens , um . And to put this in comparison , gemini 1 had 32,000 . But this is not like a small advancement .
This is really really big . And ChatGPT 4 does 128,000 . And Cloud 3 has 200,000 . So this is a pretty significant update . I haven't actually used Gemini a whole lot . Have either of you guys messed around with it much ?
Not at all actually .
No , not really all actually . No , not really . I've been , sadly , mostly on GPT , gpt-4 , and I think we use some of the cloud models occasionally as well with some of the tooling we use . But yeah , I haven't really touched Gemini , but I've been hearing a lot of good things .
Yeah , I was going to say cloud . I mean , I think most people I know use chat , gpt , but cloud I hear a lot about as well , that people are using .
I think it's going to matter what the tooling is Like . Just to point that out , right , most of us , most of us that are doing work , are going to be using a tool that uses some , some backend , like whether it be Gemini .
So I think maybe let's see what Gemini comes out with , or what it gets rolled into , and then we'll see how much we end up using it . Yeah for sure .
So with this announcement , they also talked a lot about AI agents , and if anyone has heard us talk about ChatGPT , this is kind of like the custom GPTs right , it's like you can make your own .
Some of the examples are customer agents , employee agents they had a whole diagram on the screen of creative agents , data agents , code agents and you build these with vertex AI , right , and they can talk to each other and then , obviously , interface with humans . So the idea is that you know , we keep joking about that agent that was made .
That was really just chat GPT for the car dealership . This is a legally binding agreement . You're going to see a lot more of this . I think .
If you hear anyone talk about AI , a lot of people are predicting that agents talking to each other and kind of breaking the functionality into smaller pieces is the way this is going to go , and then there'll be something that ties the agents together . So I think this is just .
AI microservices , yeah right .
Yeah , exactly , absolutely .
Hey , that episode we did with Dewan he was talking about this , wasn't he ? About the roles , like the chat , the GPT , custom GPT with roles and stuff , which we thought was pretty interesting at the time . I guess this is just another implementation of that , then .
Yeah , that . And also , if you guys remember the , when I talked about the NVIDIA conference that happened recently , I brought up the NIMS . That's right , I don't know what they're called like inference models or something that's right which are basically microservice AI agents , right ? So everyone has one big , big surprise for all this AI stuff , right ?
Yeah , real , real quick . Palo Alto , I guess , announced during this that Gemini is their AI engine of choice . Right , for all their threat detection . You know products that they have . It's interesting because I guess the Google Cloud firewall that they use uses Palo Alto's threat detection as well .
It's a partnership already , so it's like how much weight do you put into both sides saying the same thing ?
I mean , they're under the covers , just like all the other ones . They're probably running some Palo code . All right , let's see .
So we got Gemini Code Assist . Chris , you added this one . Is there anything particular you want to point out about this one ?
Yeah , I think on the service , gemini Code Assist is obviously a competitor with Microsoft Copilot , if you're familiar with Copilot . So it's basically just an integrated AI tool that will help with generating code . You know , microsoft Copilot has been insanely successful .
I think that has actually led to a lot of integration in Azure as well , so it's kind of helped them out a whole hog , whereas Gemini Code Assist is it's not really new . It was , if you remember , duet AI , which was what they had for developers . I think they just put that out . That's not even that old . I think it was like two to three years old .
That's Google for you . Yeah , and then you know , I mean to be fair . They did say that it was eventually going to be integrated with the Gemini Pro models once they were available . So it's not like the writing on the wall was there . I guess to some context .
But the one thing that stood out here , I think , is obviously , with Gemini CodeAssist being launched and the new Gemini 1.5 Pro having that 1 million token limit We've heard , you know now this allows for potentially for developers to put their entire code base into these models and you know .
So the model will have full context about their code base and can offer improvements , you know , optimizations and things like that .
Um predictive help you out with predictive stuff as well . Yeah .
Yeah , that too , um , I mean , which don't get me wrong Sounds cool , sounds , sounds like it could be really helpful to have someone with with all that context . Um , especially someone on the development side like Google .
I can't help but think on , like me , as a like in my personal life , how much I try to avoid giving my data to Google and how impossible it is to do sometimes . But , like I don't know , do we see this being like , like , is that a trustworthy thing at this point to give your entire code base to Google , or how do you guys feel about it ?
I think that definitely depends on who you are right . It's funny because right before this , we were talking about some of the sessions from Google Cloud Next and how I haven't even watched them , because you have to actually register for an account just to watch these sessions , right ?
You're like you have to actually register for an account just to watch these sessions , right ? So I don't know , it's not . This isn't exclusive to Gemini , though , right , like everybody's worried about putting data into these models .
So and losing it in the model , never pull it out again .
It's competitive advantage , right , Like if this , now people can pull from this like you're giving away your a lot of your competitive advantages as well . So depends on who you are , but there's a lot of considerations when it comes to this stuff .
A lot of organizations are , successfully or unsuccessfully , trying to build their own private models for this exact reason , either for compliance reasons or because they literally do not want to hand out their proprietary code to some third party . Even whether or not you trust Google or any of them , or Microsoft or AWS or whatever yeah , I don't know .
I think that's a big deal .
Like you said , it's a bigger question for just putting any of your code into any of these models at any given time , because , while you are paying them for these services , at the same time I'm sure they would gladly take any opportunity to you know , potentially , if you're competing with them in any regard , to build an app that's better than yours , that does
what yours does plus more . And then , if they have your code , you know it's hard not to be skeptical in that situation , so to say you know what's the lock ?
what's keeping them from looking at it is the real question right at the end of the day or using it right .
Like you remember , when uh copilot was was announced , everyone was freaking out because it was basically just scraping right that .
That's right . It was going to be the same thing , yeah , oh yeah .
So the next one . This is kind of cool . So so we talked a lot about this a while back , probably , like you were talking about this before .
It was like a year ago .
Yeah , yeah , so custom silicone right . So far , aws , as far as we can all remember is is the only one that kind of has their own custom silicon .
Right , that's those um . I always mix up graviton or nitro it's okay .
Nitro is the hypervisor right . Yeah , yeah . So google announced axion right , which is their own custom arm based cpu . I don't know if you guys have any thoughts here . Really , I mean , I this was kind of expected . We , you know , we talked about and predicted that I think all the csps are going to start coming out with their own .
I assume microsoft is probably this year will announce their own . Uh , maybe next year at the latest , but I guarantee that all of them are working on it . You guys got any thoughts ?
I'm kind of just surprised it took this long . Um , if I'm being honest , um , I mean , I don't , I don't , I mean to be fair , I'm not looking for it , but I'm not hearing these like insane success stories related to graviton . Obviously it's it . It probably does help , uh , in terms of cost and things like that .
But , um , you know , I mean , like you said , this was expected , right . Um , we , we know the other CSPs want to compete with Graviton , so this is kind of the first iteration of that . Yeah , microsoft , I know that Microsoft and IBM and everybody's been talking about how they are working on it , but no announcements yet .
So we'll have to see where that goes Well , and have those costs been passed on ?
¶ Cloud Service Mesh and Performance Optimization
Yeah , I kind of wonder too . Is it really just cost savings for the CSPs , or is it actually being passed on to the customer ? It doesn't seem like there's been any difference .
I don't think so We'd have to go pull up the AWS pricing calculator , but I believe Graviton is being sold as essentially higher performance which it is . Because it's written to take advantage of AWS , their whole what do you call it ? Hypervisor , stack and custom everything Same with this . So I don't think it's actually cheaper to the customer .
I think it's probably cheaper for them to run because it's more optimized , and because it's more optimized , they can get better performance out of it . So , yeah , I think it's probably the case here too , although I didn't see any and there's no pricing in this article we're looking at right .
Yeah , it's kind of funny . It's almost like double dipping right , Because they cut costs by creating their own custom silicone and then they also oh , there's a side performance . So , let's raise prices as well .
Right , I mean we get . Yeah , like , definitely there's . There's definitely lots of customers that ask uh , hey , you know Graviton , are you going to support it ? Is it useful , you know ? Are you going to be using it ? So there are definitely people that notice and and want to take advantage of it . But , uh , there's , there's definitely still some .
Uh , what's the word I'm looking for ? There's some limitations . Uh , like your , your app has to support it . If you're at that level , at the instruction level , All right , so we got Google Cloud .
Assist was another thing announced . This is funny because I don't remember which episode it was , but Tim , you were going off on Amazon .
Q you got announced .
I mean , there were so many limitations , it was super expensive . I think there was only 20 questions .
It had that whole Yogi on the mountain thing where you can only ask it a few questions a day , or something and I think if you closed out of the window and came back , it would like it didn't save anything . Yeah , it didn't save anything .
So I don't know , there's not . I don't think there's a whole lot of details in this announcement , but I really hope it's better than when Q was launched .
To be fair , I haven't gone back to Q and looked at it . Maybe that was like a first blush and it's better now . I doubt it , since I haven't seen or heard anyone using it . So if you're out there and you're using Q , we would love to know is it good , do you like it , is it better ?
How many questions do you waste every day getting to what you care about ? Right ?
is it ? It's more like is it worth the money ? Because it's it was obviously priced higher than I think a lot of us expected and it's you know if you're paying for it .
I'm really curious if it's worth it well for what you got specifically and not just that was it expensive , it was like , for what you got at that time . So yeah , and Google gets a free pass on this specific episode , simply because Cloud Assist just says essentially what it does but gives no detail , like the Amazon one did .
I was going to say I will say , of the few presentations that I did watch specifically around cloud networking , like they had this one about what's new with cloud networking and GCP , and it seems like a lot of the focus was on using Google Cloud Assist to look at things like capacity limits , latency , but directly tied to your applications , not looking with kind
of these generic terms , but you could literally ask it , for you know , hey , what is ? What is latency to my particular application ? Look like in these four regions and it would give you like a graph that you could interact with .
It was it was it was actually pretty impressive , um , which I think the , the actual , the actual , uh , the way they do the graphing on the backend is is some other proprietary thing that they use , and I can't remember what it's called . It's escaping me right now , but yeah , it's , it's , it's pretty cool .
Pretty cool . Yeah , that sounds really cool . You know all this , the graphing stuff , I think right If their focus is on devs , right , a lot of the networking stuff they kind of want abstracted . So I guess it would make sense if it's got a focus on the networking piece .
Um , and I do remember I feel like parts of the article were talking about how it can kind of help you build infrastructure , um or it did say gather , you know , like what you have , it kind of like helps you optimize it or figure out what could run well on it . That's an interesting angle that I don't think q was taking .
Yeah , I think that's right . And yeah , the article did specifically mention you could actually build resources with the cloud assists and also that you could be like , hey , how do I optimize my CPU or something ? It could go figure out that information for you . Yeah , right .
So we'll see . Yeah , I mean , the tools are only going to get better , right . So even if it's not great now , I assume in the future there'll be . Yeah , we can only hope . Okay . Cloud service mesh this is one . When I saw it announced I was very excited , right Like the day I saw this blog come out I think I reached out to Yvonne I was excited .
I was like , hey , you guys are announcing some pretty cool stuff and this is a cool product . I wrote down that this is a VPC Lattice competitor . After kind of doing some more digging , I'm not sure that's what it is , because , really . So let me I'll just real quick , I'll explain what this is . It combines traffic .
So there's two existing products Traffic Director and Anthos Service Mesh , and Traffic Director is basically a managed control plane for your service mesh , and Anthos Service Mesh is a managed Istio-based service mesh .
So cloud service mesh I know a lot of alliteration here , right , but cloud service mesh integrates both of these into one managed control and data plane service mesh . Right , and it's specifically they call out . It manages a data plane of Envoy sidecars , which is the proxy that pretty much all service mesh companies use .
Like Linkerd wrote their own , so I guess they wouldn't apply . Cilium doesn't really use sidecars , but everybody else uses Envoy . It's a very popular sidecar proxy . And then also like this like Lattice right , how Lattice goes across VMs and containers .
And like Kubernetes , endpoints and Lambda right , like this also goes across VMs , containers , cloud Run , all the different pieces . So it's I don't know . It's kind of similar to Lattice , but in a different way . What do you guys think here ?
Well , it doesn't have the specific remember , lattice has specific focus , right ? Like the only TCP-based application , like there's certain limitations of the way the Lattice is built . So , yeah , definitely different , but I can see functionally similar , which you know .
It's funny because when we did that in an episode with Alexandra and Justin , they were like , don't call it a service mesh , but you know , a spade is a bit of a spade , right ? So yeah , the fact that it looks similar is not surprising .
Yeah , lattice seems to be . I hate to say that it's a little bit more focused on re-architecting because I don't feel like your architecture changes that much . But , like some of the core components in VPC , lattice are much more . You have to be fully integrated with AWS using that . This seems like it could maintain some portability , but I mean that could be .
I don't know that could be short-sighted . It's kind of hard to tell . I'm not as well-versed in this . So maybe , alex , you know , if that's , if it sounds that way at all .
Yeah , I mean , that's a good point and I do think there is some some validity to that , right , because , like I was talking about Envoy , right , sidecar proxies , you can deploy them anywhere , right , it doesn't matter , it's not like that's a GCP thing . But also , of course , all the cloud providers have their own Kubernetes offerings .
So if you're using , like GKE I don't know all the terms for , like what they call it when it's in a different cloud , right , Like how they have EKS anywhere , and I don't know what the Google name for that is , but I imagine that this would extend anywhere you wanted to place it . So that is a really good point .
And , yeah , it's different from VPC Lattice in that way because , yeah , you are tied in specifically to AWS . And , real quick , justin , if Justin is listening , he can tell me not to call it a service mesh all he wants .
But that's my first thought and I know it's like a service mesh on steroids , but that's how I've always looked at it Shots fired , all right . So I got another one . This is just a super quick little fact that I saw Google distributed cloud hosted . That's a mouthful , but this is basically their secret and top secret workload approved cloud hosted .
That's a mouthful , but this is basically their secret and top secret workload approved cloud offering . So this is comparable to AWS , govcloud and MAG , which is Microsoft Azure for government . This is one like . I feel like this should have came out a long time ago . I don't hear many organizations using GCP for any kind of work , so it'll be interesting .
They're behind on this , I guess , is my point . It'll be interesting to see what the uptake looks like now that this is available .
Yeah , I mean , I don't know about you , tim . I don't know if you're seeing a lot of GovCloud customers . It seems like I don't hear about it . I mean , obviously you have certain customers that are all in on it and aren't going to use any cloud services unless they're integrated there .
But yeah , it's like , yeah , they're late , but obviously there's so much more of the market to to tackle than this so I'm not faulting them for being late with this . At least you know now they can . They can , you know , scrape up some of that as well , but yeah .
I was just saying that . I think that the question is going to be whether or not the they made significant improvements on the things that the federal government would appreciate about a government entity Right . So that's the question is is this better ? The federal government would appreciate about a government entity , right . So that's the question Is this better ?
I mean , they're all FIPS compliant , for example , but does this have extra stuff that the government's asked for that maybe AWS doesn't have because it's oldest or something ?
Well , and it's a fair point , chris , that most agencies probably none are fully migrated to any kind of GovCloud anyway . So I guess , right , maybe they're not late , right , Maybe they're perfectly on time .
Okay , so now we're going to get into some ones that are going to be a little bit confusing as we call out these names , right , and we'll get to that in a minute , but let's start with cross-cloud network .
Okay , so I want to preface all this by saying this is our understanding of what these products mean , right , and try to do a bunch of digging into this . Some of these names are very confusing and redundant , right ?
So cross-cloud network and subsets of other ones , yeah .
Yeah , exactly , and they're tied together and some pieces are not in one product but they're in another . But cross-cloud network is the overall name for , I guess , the platform GCP has created for networking right .
¶ Cross-Cloud Interconnect and Network Integration
So this includes their next-generation firewall cross-cloud interconnect , which is how they connect to other clouds and then private service connect right , and this just allows , like it's .
It's basically like Private Link . Yeah , PSC is like Private Link .
Okay , that makes sense . So what do you guys think about this one first , before we dive in any deeper so , yeah , I mean , I went through the article .
This includes not just CCI , the cross-cloud interconnect , but what is it ? They're talking about ? How you can connect SASE stuff to it , sd-wan .
When I read cross-cloud network as a platform , what comes to mind is either like a cloud WAN or a V-WAN , depending on how you want to talk about it where it's going to centralize connectivity from what they call hybrid and multi-cloud sources . In fact , that's what it says in the diagram hybrid and multi-cloud sources .
It's a little different than both VWAN and Cloud WAN , but it's also just so confusing the way they've stacked this all up . Anyway , we'll get into that in a minute .
Yeah , pretty much . I mean , I think that is a good point that it feels like a mix between , maybe , cloudwin and VWin , whereas it's kind of all contained in this thing . You get all these service enablements within CCN , or cross-cloud connect as we're calling it , but it also has these interconnect points to integrate .
I think that's another thing that's really important here is that one of the main focuses is integrating other CSPs where you don't see that a lot with any of these other , you know , kind of full-fledged networking products from the other CSPs .
There's a there's a strong focus on interconnectivity with outside organizations , whether it be on-prem other CSPs and things like SD-WAN . But then it's also confusing because some of that stuff was also kind of offered through NCC and I guess that was kind of the confusing piece is like this is not NCC . It was pretty clear that they're not the same thing .
NCC being Network Connectivity Center , which is a front-end for Google Cloud router where you can configure it . That's what NCC is for , sure .
Yeah , yeah , right , and it's weird because a lot of the documentation on NCC makes that seem kind of more like VWAN , right , where it's like a managed hub . So there's so many similarities . It's just weird . I wish they did a better job of kind of clarifying , but that's just me .
A good thing , like Chris pointed out , is that who else is integrating different clouds , right ?
Nobody's talking about it at least . Yeah , no , nobody's talking about it . Well , oracle , no , oracle's talking about it to be fair .
Well , that's true . That's true . Oracle definitely does Oracle ?
Oracle's talking about it , but I think , once we get into the weeds about cross-cloud interconnectors , cci , while , yes , there's a strong focus on interconnectivity with other clouds , I will be honest , when I read how CCI is done , it's a lot more elementary than I thought it would be . But it is a lot more elementary than I thought it would be .
But I mean , it is a start , I mean , and it's , let's get into it .
Does anyone want to talk about ?
So CCI .
Cloud Interconnect or Cross-Cloud Interconnect , is basically Google's offering saying that you can use the private Google backbone to interconnect with other CSPs , being predominantly AWS , azure , oracle . Even AliCloud is listed on there .
When you kind of look into the reference architecture diagrams , it just looks like they have probably dedicated 10 gig and 100 gig connections , potentially in some kind of meet me room , with these other csps , and it really just looks like when you configure these through cci , there's a dedicated google cloud router in the co-location facility that can terminate the
layer 3 connectivity . But it also sounds like there's no orchestration at all on the remote end . So you're still deploying direct connect . You're still deploying Direct Connect , you're still deploying ExpressRoute in the exact same fashion you would with those , which totally makes sense .
I understand that Google isn't in bed enough with these other CSPs to make this a fully orchestrated product between the two , but at the same time it's like I could have picked up the phone and done that myself .
I didn't need you , google , able to help me with this part .
Yeah , I mean , in one of the presentations I watched there was a customer success story from I think Citadel is who it was where , you know , they had a migration project where they had to get something done with cross-cloud connectivity , data transfer that needed to take place within you know , the order of two months or something like that , and it was a ton of
data that needed like like 10 terabytes of connectivity per hour and things like that . So it was . It was a lot of stuff that needed to happen and the orchestration of that you know , doing it on-prem was , you know they had to orchestrate , you know , enough network infrastructure to own the capacity , enough circuits to own the capacity .
Doing that all themselves would have been a much larger burden than just using something like cci .
So I see the opportunity , but it's like I don't know the , the , the consumption of it just seems I don't know how much , I don't know how it's priced either , like , if it's , if it's , astronomical , then that would not make sense to me , but I , I , I , I would imagine that it's not . Um , yeah , it left a little bit to be desired , if I'm being honest .
Yeah , I mean let's be clear .
What the CCI offering is is Google will set up a Google router in Google and connect a Google circuit over to the other , so over to say , an Amazon router where you , as the customer , went to Amazon , bought the Direct Connect , connected it , and then the Google Cloud router is essentially just replacing your on-prem router that would peer over the Direct Connect
with , say , a VGW or something . So instead of your router that you would be connecting , now it's a Google Cloud router . That's the extent of the integration . Again , good start , but I'm with you , Chris , If that costs a lot of money to do , I could have just picked up the phone and done that myself . I didn't need it .
Yeah , I mean , like you guys are saying , it really just takes away that middleman , right , then you don't have to worry about terminating on either end of your own gear and renting out like Colo it costs like , put you know , rent out a room in a colo facility and then bring those connections both into your own gear yeah cost is a big deal here to think
about yeah , if they charge an arm and a leg to do that which you're already paying , by the way , for the circuit from google and for the circuit from , you know , the other cloud , that notwithstanding the service itself .
Um , so that's already a decent amount of money , right ? If they're charging you a mint on top of that just to let one cable go over to the other provider router so you can do the peering you were going to have to do anyway , that would not fly with me .
I feel like it would make more sense if there was some incentive to do this , as far as maybe the data transfer rates were less or something like that , but it would have to be .
It's pretty underwhelming .
honestly , it's underwhelming but I understand that AWS and Azure , Ali , OCI aren't going to be in cahoots with this product and offer this kind of seamless thing .
No , they're not excited to jump into this .
Well , you need a third party . You need a third party that's agnostic to all of them , that will be that middle person and connect those clouds and just be that .
So you don't have to build your own infrastructure in that colo , because that's what's missing from this is that which already exists uh to a degree , and it's equinix and it's megaport and it's all these people that are doing things like that right . So I mean I don't .
I don't , unless you have this like significant orchestration piece from the control plane side that is built within cross cloud connect . That is a benefit to the actual end customer . I don't see why you don't just do this with a third-party uh provider like equinix or something like that well , that's the question .
I have right , because a lot of the diagrams , um , it's hard to grasp exactly what they're kind of , what story they're trying to tell with some of these diagrams , but the cross-cloud network stuff , they they at least , are talking about it and making it seem like they want to bring in those into this overall platform Like a hub thing , yeah .
Yeah , it's kind of like they're at the beginnings of integrating everything into one platform that you can manage everything from . So if it evolves to that , I'm all for this . Like , yeah , this is super early stages , but I hopefully it does right , Hopefully it does evolve into something more where you can do stuff with it .
Yeah , I mean there are options for integrated security with like this , you know , the offering of perimeter protection with the secure services edge . That seems cool . That's something that is referenced in CrossCloud Network . That does look cool and , I'd say , probably a little bit more advanced than maybe other offerings at the current time . We'll see .
Yeah , I think so , but they're integrating things like Fortinet , palo . Even Broadcom was on there too .
Yeah , Symantec . Symantec by Broadcom Makes sense , everything by Broadcom , everything by Broadcom , everything by Broadcom , carl's .
Jr , brought to you by Carl's Jr Cross-Cloud Network by .
Broadcom . Yeah , and to be clear , cci , clark's College of the Interconnect , is a subset of Cross-Cloud Network , so that's just like one more connectivity option .
To be clear .
Well , you have to be , because this shit is like the naming is so clear . Yeah , yeah , well , you have to be , because this shit is like the naming is so awful . Anyway .
Yeah , they got to figure something out , the other one , like we've talked about . So Network Connectivity Center and CrossCloud Network seem to kind of share a lot of similarities .
But also they have Google Distributed Cloud and I just like someone needs to just clean up some naming and tell the story better of what these things actually are , because I thought I knew what distributed cloud was . I thought it was some kind of like subset with Anthos and you know , being able to spread out like , manage Kubernetes , clusters everywhere .
But I'm not even sure that that's right anymore .
I don't know . I mean , they can do as Google does and just kill it off like lean into something . True , you know , make it easy for us All .
Right , I got two last quick ones and then we'll wrap it up .
¶ Discussion on Google Cloud Networking Updates
So there was model as a service endpoint , right , and this is basically what it sounds like . This is allowing you to provide an AI model to customers , right ? Yeah , I don't know if this is similar to what AWS does with Bedrock , is it ? I mean , I don't know .
If it's an endpoint , probably not , yeah , no , yeah , it's probably not right . It's probably using Private Service Connect and they have like a load balancer on their side or whoever's offering the model would have like a load balancer on their side and you offer it as a service .
I don't forget what the Google version of it is , but in AWS you have a marketplace . I'm sure Google has a similar marketplace .
I think Bedrock is the marketplace right .
Oh , maybe for that it is yeah , maybe Right , okay .
And then we got this one we've talked about actually a couple of times , but the Google Cloud VMware Engine it's just that they've on their end has announced the VMware Cloud Foundation license portability , which is kind of cool , right ?
If you're all in on VMware which means now you're all in on VMware Cloud Foundation you are able to have that portability between running your VMs and containers . I guess if you're using Tanzu .
Or the clouds , yeah .
Migrating to GCP . I don't know what that migration piece looks like either , cause I heard that HCX isn't really a thing anymore , so I don't know .
They're just going to stretch VLANs across the do V motion .
Yeah , I mean this . Yeah , this is . This is cool . I guess , if you're , if you're all in on VMware , but at that , at this point , I think , if you're still all in on VMware and remaining all in on VMware , you probably have bigger fish to fry .
This , this , this was probably not your , your problem to solve at that point . Yeah All right ?
Well , that is everything we wanted to cover . I just you guys got any last quick thoughts before we wrap up , or you good ?
No , I think , you know , I think it was we wanted to do this episode , cause I do feel like we obviously we focus on the top two players in this space a lot being AWS and Azure , and you know , I feel like we we should give some love to GCP or sorry , google cloud . It's not GCP anymore .
Is that right ? It's not GCP anymore . That is correct . They renamed it to Google Cloud because they love naming and renaming . Maybe this is why they're third .
We'll see . But yeah , I think it was nice to give them some love . I mean , obviously they have some really cool technology . Some of it is playing catch up , which they've had to do , and you know , ai has probably been quite the disruptor for them . But it's nice to see some of the stuff coming in , like cross cloud network .
It's just it's hard to maintain the knowledge base around all these things when they're so closely named and kind of do the same thing . But you know , we've been doing it with AWS for a while . Why not ? Why not one more ?
Yeah , for sure , tim , you got anything .
Yeah , I , yeah , I kind of feel the same way . We definitely , as multi-cloud networking folks . We don't spend a lot of time on GCP because GCP networking is so alien to traditional networking , but we need to put more effort in on it .
I think that there's still a lot of value here and it looks like they're moving in the right direction with some of these new offers , even how goofy some of them might be . I do wish they'd get their naming worked out . But yeah , I'm excited to see what's coming next .
It looks like they're paying a little bit more attention to networking , and that's already important .
Yeah , for those that aren't familiar with in Google , I mean Tim . Yeah , for those that aren't familiar with in google , I mean tim . You tell me if you agree . But when you learn networking in google , uh , or in google cloud specifically , it feels like they did a lot of work to make it something that is not considered by anyone .
Right , it's like you want to be completely obvious to the developer .
Yeah , you don't have to care about it whatsoever .
Yeah , please don't do auto mode , if you ever . If you're a , you don't even have to IP things .
Please don't do auto mode . If you're a developer , you don't care , but yeah , so it's nice to see them make some strides like this kind of refocusing back on some kind of core components of networking .
So yeah , we'll see where it goes . Yeah , agreed , I think it is exciting , like you guys said . I'll just echo what you guys both said that it's nice to be able to focus on not just Azure and AWS .
Right , gcp and OCI are definitely still heavily investing and making advancements and it's good to see there was a lot of networking stuff at this conference , so it's good to see that they're listening and evolving .
Does Oracle do one of these ? We should probably look and see if OCI actually does one of these .
I know there's some conference . They do that's later in the year . I can't remember what it's called right now .
We'll keep an eye out for that one and see if we can get somebody from Oracle to talk about it with us .
All right , well , with that , let's go ahead and wrap up . So , um , as usual , if you know , if you like the episode , uh , please share it around . Let all your friends , let your family know about it , your mom , your grandparents yeah , mom , dad , you know grandma yeah , mom dad , grandma , everybody , um , like , like and subscribe , all the things .
I think we're on every platform , as chris likes to say uh , except instagram , right , we're not . We're not on instagram , I Instagram right .
We're not on Instagram . We're not on Truth Social .
Oh , that's true . That's true . We're on both X and Twitter , though .
Yeah , both X and Twitter yeah .
All right , and then , yeah , as always , feedback is always welcome , right ? Please reach out to us . Send us an email , cablestoclouds at gmailcom . Always welcome , right ? Please reach out to us . Send us an email , cablestoclouds at gmailcom . We're always open to have more conversations and talk about things that we might not be thinking about much .
So thanks for listening and we will catch you guys next time . Hi everyone , it's Alex and this has been the Cables to Clouds podcast . Thanks for tuning in today . If you enjoyed our show , please subscribe to us in your favorite podcatcher , as well as subscribe and turn on notifications for our YouTube channel to be notified of all of our new episodes .
Follow us on socials at Cables to Clouds . You can also visit our website for all of the show notes at Cables to Clouds dot com . Thanks again for listening and see you next time .