They got rid of that motto a while ago, by the way. It was officially removed from their company statement or whatever it was. Google...err, Bing it, if you prefer.
Makes sense, because their business model has moved sharply towards the "evil, predatory" side of the spectrum.
RolandDeschain wrote:Massive antitrust, whoop-de-do. Did you think that would end Microsoft back in the day, too? Apple's had their share as well, though their insane amount of money spent on marketing has helped protect them in the public eye.
This will be the biggest antitrust case in history, dwarfed only by the one that Amazon will eventually be hit by. At present, it's really looking like both orgs are going to get hit with antitrust action in the USA and EU and it's just a matter of time. Given Google's current trajectory, I think antitrust will be the least of their concerns. There are prosecutors talking about charging them with treason for providing so much tech to the Chinese government, and that's a case I'd support blindly. Given their cozy relationship with the Chinese government, it's amazing to me that you're still such a huge fanboy of theirs. But who cares about the wholesale subjugation of billions, amirite?
RolandDeschain wrote:1) Nobody scales infrastructure like Google does.
Debatable. When I was at Amazon, we scaled the network to multiple millions of hosts, far beyond anything Google was doing at the time. Either way, autoscaled infrastructure is a great way to waste money when you lack the expertise to properly forecast and plan your capacity needs. The OpEx approach makes for easy accounting but for large enterprises it ultimately costs a lot more than the CapEx approach.
RolandDeschain wrote:2) Nobody knows how to search data like Google does. Amazon's not even remotely close. Nobody is. Restrict their Search revenue and they'll branch out more.
This isn't some special thing that only Google does. Quickly searching massive data sets has been a solved problem for a decade or more. The issue is with user perception, namely getting users to consider other search options. That's a difficult thing to do when people have been trained like Pavlovian dogs to go to google.com the moment they open a browser. The only thing Google has managed to do that is special in this space, is monetize search results.
RolandDeschain wrote:3) Google Cloud Platform is rapidly growing. They don't offer anywhere near as many random other services as AWS does, but what they do offer, they do better.
Also debatable. I'm not really much of a cloud services evangelist as it's cheaper for large organizations to host their own infrastructure.
RolandDeschain wrote:Smokin', maybe you can tell me why GCP instances in Oregon have lower latency to EC2 instances in Ireland over the public Internet than Oregon EC2 instances do over AWS' own internal global network, lol.
Without data like ping results, traceroute, hardware the GCP instances run on, hardware the EC2 instances run on, and so on, it's pretty much impossible for me to say exactly why. If you are simply looking at the round trip times for the ping command, you are not getting concise enough data to declare that there's a significant delta between the two services for application latency. I'm also having a hard time understanding your use case. Why would you want to transport so much data, so far, over the open internet, when latency is a concern? Dumb. That sort of thinking is why my previous employer cratered six months after I left. Bear in mind that cloud services are inherently shared infrastructure, so an apples-to-apples comparison is never going to be fully attainable.
Also, iperf throughput testing shows GCP absolutely MURDERING AWS. 1.6Gbps from GCP Oregon to GCP London.
I was fortunate enough to have a client want me to do some basic network testing (they are looking at putting some of their infrastructure in one of the major public cloud providers to get faster application latency over their global networks to offices in other countries so they could kill their ludicrously expensive SD-WAN service), and the initial results are embarrassing. Obviously, the testing I did was just quick & dirty; but it's still pretty damned revealing.
Only 1.6Gb/s? What's wrong with your endpoints (aside from the fact they're cloud based)? Something tells me you're not tuning your TCP stacks to accommodate the long, fat network in use. Research the "long, fat network problem".
It doesn't sound like your client learned their lesson with IAAS offerings. Can't say I'm surprised. That $h!t is like the Hotel California.
RolandDeschain wrote:Not to mention, managing almost anything in GCP is faster, easier, and more intuitive than the AWS "designed by engineers in their moms' basements" web GUI. AWS needs to steal GCP's web GUI UX in a big way. AWS enjoys their large market share for reasons that are not long-term sustainable. The trend of a good portion of the growth that GCP and Azure enjoy coming out of the AWS pie will continue. Especially as time goes on, and fewer old IT director guys stuck in their ways make decisions along the lines of "let's go with that newfangled AWS cloud thing", lol.
You seem to think because of my history at AWS that this is a matter of personal pride for me. I assure you, I really don't care. Over dependence upon cloud services is dumb, especially for large organizations that can afford to develop their own infrastructure, which is why many organizations are doing just that.
RolandDeschain wrote:The TL;DR version is: Google's expertise and inventiveness that got them to be the Search & Ad king in the first place can be applied to other sectors, and if you neuter the advertising portion, you won't be neutering Google. Blue cells are GCP wins, orange cells are AWS wins.
Actually, I think it's more a matter of them being early to market with their search product, and the massive infusion of cash they received from the CIA via their cutout In-Q-Tel.