GCP advantages & benefits over AWS/Azure
Google Cloud Platform or GCP has raised the bar for other cloud service providers since last 2 odd years. What was increasing looking like a two horse race between AWS and Azure, has now changed due to increased participation from GCP. A CxO had a binary choice until very recently, with GCP deciding to aggressively move forward on products and services - has led to a 3rd alternative. And the first question they face is - What are the advantages of GCP? Why would one choose GCP over AWS or Azure?
It is important to note that Why AWS over Azure or Azure over AWS was the starting point only few years ago. An enterprise just went ahead with one of the two.
With Cloud computing, an enterprise could meet today’s demand and lay a foundation that allows it to leapfrog ahead of competition in future by utilising cutting edge and innovative products, dynamic pricing models, X-as-a-service model and increased agility.
While these are common benefits across a cloud vendor, GCP has set itself apart from others in few broad categories and we will examine each in a bit of detail in this post.
The main benefits/advantages are:
- Tailored Compute
- Superior Fiberoptic Network
- Hybrid Cloud Products - Anthos, Kubernetes
- Machine Learning and AI - Tensorflow
If you take a look at the broad Big Data/Hadoop market, you would see a range of jargons, tools, frameworks - Map/Reduce, HBase, Apache Drill, Parquet Format and so on, the list goes on and on. Then you have other open source initiatives like HTTP/2, gRPC, Protocol Buffers; machine learning products like Tensorflow, contain management solutions like Kubernetes, Hybrid cloud solution like Anthos, even processes and practices like SRE - all these diverse tools/frameworks has one thing in common - Google and its open source contributions.
Your chances of getting a cutting edge and innovative product before the competition is much higher with Google - GCP.
With GCP, one could got with any combination of compute and memory while the fast disks and networking with any instance. One need to go with a "Network Optimised Instance" and discover CPU and/or RAM shortages in production. Instance choices are simpler and not segmented based on Compute, Memory, Storage and Networking. A user has full flexibility to change any one of the four features of an instance - Custom RAM and CPU combination with configurable IPOS, disk size, bandwidth and networking.
Simply not possible with other providers.
Superior Fiberoptic Network
Google uses its own fiberoptic network with provides ultra low latency and has been widely documented by 3rd parties as well. It also uses a global SDN (Software Defined Network) which offers incredible performance specially for Microservices and big data work loads.
This private network means - an enterprises workload travel on a dedicated superhighway bypassing the congested internet lanes altogether.
Hybrid Cloud Products - Anthos, Kubernetes
No other provider is investing and innovating in products that gives an enterprise customer option to avoid vendor lock-in.
With Kubernetes, it allows applications to be written and managed in a machine/os/vendor agnostic way.
At the same time, with Anthos, it allows those application to be run in any cloud platform.
As per Gartner, the future is hybrid and only GCP seems to be making it possible.
Machine Learning and AI - Tensorflow
This ties into Innovation as well. If 2005-2014 was an era of Big Data, there is no doubt that 2015-2024 would be an era of Cloud Computing - where it would be adopted widely. The future would be an era of Machine learning and AI where its usage would become commodity, it would be easy to use and every enterprise - big and small - would have some form of ML/AI.
GCP stands to deliver on that future as well. With Tensorflow, it has made ML a framework/tool that any one could plug into.
Finally the cost, independent comparisons put GCP to be about 50% cheaper than AWS.
There are few points worth mentioning on pricing:
- Per second billing for all services.
- GCP is cheaper by 30-50% from competitors while giving better performance.
- GCP's strategy for competing on price is to offer discounts for sustained use, for customised use, and for committed use automatically.
- GCP's Preemptible VMs or Spot Instances are a fixed flat 80% off of list price. There is no spot market to deal with. This fixed price allows an enterprise to project costs consistently.
While there is no debate around which cloud provider gives more services and tools - there is a question around which provider gives tools and services that are different and are tailored to global scale.
GCP provides one such tool that no other provider has -- it is not Kubernetes and Anthos. It is Spanner. It is best of both - SQL and No-SQL worlds; it has global scale and strong consistency guarantees.
Why GCP again?
Cloud computing is not a question of if but when. And the answer to that when would define a company - it would make or break an enterprise. IT is no longer a cost centre, it is an enabler that allows enterprises to forge ahead of competition. Cloud enables rapid changes in enterprises and no one had been able to handle rapid changes better than google.
The differentiator going forward for enterprises is going to be the speed and agility with which they can achieve a goal - cloud enables that and google cloud augments that with cutting edge technology and innovation.