- [Instructor] As a working Cloud Architect, I design and implement solutions in a number of public vendor Clouds, and those include GCP, Amazon Web Services, Microsoft Azure, even IBM Bluemix. So I get to work across these different vendors, and I get a look at what I called the "personality" of each of the vendor services. So for GCP, this is what I see. The good, in terms of Google Cloud Platform services, are the services are fast and efficient.
There are frequent improvements to those services, because of the engineering culture at Google. Very, very commonly, if there is a weakness in the service, Google will hear from their customers, and they will make an update to that service in months or even weeks. So you see the services improve very quickly. As mentioned in a previous movie, the default with the GCP services, the majority of them, is to autoscale, and autoscale massively. So I've run into other public vendor Clouds with these service limits where only a certain amount of data can be stored, only a certain amount of compute can occur.
I really don't run into that. And some of my highest volume customers, I work in the southern California area very frequently, so social media, ad tech, really big volumes of data, they will often prefer the Google Cloud, because of these aspects of strength in the offering. Overall, they're really good value for powerful and scalable compute. Now, that being said, they're not perfect. So they've got some areas that are not strengths. In fact, some of these I would say are even weaknesses.
Because Google itself is an engineering culture, their products are really designed kind of with engineering focus. So what does that mean? Specifically it means, very frequently, like everybody else, when they introduce a new service, they start by introducing it, and it's only accessible via scripting tool. It's not acceptable via the console, usually, first, so you have to use script. Sometimes, you have to use an API or an SDK. You have to program against the service, so that does that restrict for some of my customers, who are maybe more of a dev-ops bend.
They're not programmers, they don't want to write code. The use of some of the services. This isn't across all services, but it is something sort of ongoing with the Google Cloud. The ease of use aspect, the click to get started, the samples and tutorials, they're more of an afterthought in this world, rather than in some of the other vendors where it's more of an end-to-end product offering. So not always easy to use, particularly if you're not a developer or a programmer. The Achilles heel for the Google Cloud, if there is one, is documentation.
Google is well aware of this, and they do care, and they're working on it. So if you do work on the Google Cloud, and you have time, and you find problems with the documentation, do let them know, and they'll fix it. For some reason, this has just been a lagging area since the inception, and just out and out, it's a problem. So when you are working with the services, work with the documentation, and if the documentation doesn't seem to work, it's probably not going to be you, it's probably that there's an improvement in the services. Now what should you do to get around this? You want to, as much as possible, be part of communities, join the product forums, get to know the Google Cloud developer expert community, who's a partner community.
Because a lot of us blog, a lot of us do code samples, a lot of us will try to correct the documentation. It's just a real challenge. Documentation is frequently not matched to the product. The next area of the Google Cloud to pay attention to is pricing surprises. This has been particularly the case around some of their data services, most notably, their query has a service offering that's called BigQuery. Although, again, they have added some pricing information around some of their services, for example around their compute engine or their virtual machines, they actually display the pricing on the console.
So they seem to be understanding that this autoscaling within Google is not an issue, because they just want to scale their own resources, but when you sell these services as commercial services and they autoscale and then the customer gets the bill, that might not be optimal. So we're going to be really looking at this in this course. We're going to be using the pricing tools, the pricing calculator, understanding how to figure out how much things are going to cost. Speaking of tools, this is another area where Google is lagging behind. Not only in their own GUI tools, although those are actually rapidly improving, last 12 months in particular, but particularly in partner integration.
And this has actually been a deal-breaker for some of my enterprise customers, and, again, this is something Google is aware of, and if this is important to you, you should give them feedback, as well. Because they are actively working on this. Again, they released their services to the Cloud at a later point than some of the people in the market, most notably Amazon, and they are working to build out their offerings, including the third-party integration that we customers really need, but they need to hear from us. So tooling, at this point, particularly third-party tooling, is really an area for them to work on.
- Google Cloud Platform benefits
- Compute services
- Database and storage services
- Data pipeline services
- Machine learning and visualization
- Networking and developer tools
- Implementation solutions
- Architecture options