Which Cloud is Best – AWS vs GCP
Recently we compared Amazon Web Services to Microsoft Azure. Those aren’t the only two providers out there even if they are the biggest, and so today we’ll be introducing Google Cloud Platform to the mix.
We’ll start by seeing how GCP compares to AWS.
Google Cloud Platform (GCP) began back in April 2008, whereas Amazon Web Service (AWS) was launched a little earlier in 2006. But does that mean AWS is automatically the better provider?
Actually there are lots of factors we need to consider before deciding. Each business has different needs to be catered for by cloud providers. So just because one organisation favours one provider that doesn’t mean they’re right for you.
To really find out whether GCP or AWS is best, let’s take a look at everything from their coverage and sustainability, to the specific services they supply. First up is one of the main reasons people turn to cloud providers – paperless file storage solutions.
File Storage and Transport Solutions
File storage solutions offered by cloud providers are increasingly popular. Not only do they provide essentially limitless storage for your data, but they increase the ease with which you can locate and share your files with others. Whether you’re a massive company or you’re just starting out, cloud based storage can be really useful. So let’s see how GCP and AWS compare.
Let’s start with Google
There are six main services to take note of when looking at GCP. Firstly, you’ve got Cloud Filestore, which is designed for applications that need a file system interface and a shared file system.
Secondly, is Cloud Storage, a high-performance object storage with four storage classes. You can move your data across these classes as you choose, depending on where performance is best optimised.
Thirdly, is Persistent Disk, a flexible block storage that offers multi-reader capabilities.
Next, we’ve got Cloud Storage for Firebase, which is used to store user-generated content.
The next one to take note of is data transfer services, which can be used to move structured and unstructured datasets into Cloud Storage, BigQuery, or Cloud Dataproc.
And finally, there’s Drive Enterprise. This provides a place for collaborating through storing, sharing, and editing files. So GCP has a good variety in terms of file storage.
But what about AWS?
Amazon has a stunningly large array of storage solutions.
First things first, there’s Amazon Elastic Block Store, which offers low-latency block storage for Amazon EC2 and helps you tune applications with the best storage capacity, performance and cost.
Amazon’s Elastic File System is a scalable storage system for AWS Cloud services. There is also the Amazon Simple Storage Service (or Amazon S3), an object storage to store and access any type of data online.
Amazon FSx for Lustre is a file system for compute-intensive workloads.
Similarly is Amazon FSx for Windows File Server, which lets you move Windows-based applications to AWS if they need file storage, with the same compatibility and features they rely on already.
Amazon S3 Glacier is an archive storage for long-term backup of any data type.
AWS Backup is pretty self explanatory – a backup storage service that can work alongside AWS Storage Gateway. This basically integrates AWS Cloud storage to your on-premises location.
And a final storage solution worth noting is Data Transfer Services. This is a variety of services used to migrate your data to and from the Cloud, helping you make the most of these services.
So you can see both GCP and AWS have a lot to offer in terms of storage space, and services to move your data into this space. We always advise you select based on what fits your needs best, however in terms of the sheer variety of services, it’s fair to say AWS wins this time.
No cloud provider currently has complete global coverage, but they’re all expanding their regions, zones, and datacenters. This is important because sometimes certain products and services are only available in specific regions.
If you’re looking to develop on a global level, you’ll also want as much coverage as possible. With a higher number of datacenters comes a larger number of places you can choose to store your data. So which of our two cloud providers has the greater coverage?
GCP has a map of its current zones and future planned zones for anyone to browse. They currently have 61 zones placed within 20 regions. But they have plans to expand into South Korea (Seoul), further into the USA (Salt Lake City), and Indonesia (Jakarta). Not every single GCP product is available in each region or zone. So it’s important to decide what you need before choosing GCP, to make sure these services are available in your region. Generally, GCP regions include the majority of their services, but it’s always worth checking beforehand.
AWS is slightly larger than GCP in terms of regions and zones it covers. But only slightly! They currently have 21 availability zones, within 21 regions. But they have announced plans to expand with 12 new availability zones and four more regions. These chosen regions are in Bahrain, Cape Town, Jakarta, and Milan. They also have a handy map for customers and potential customers to view each area they cover. Again, it’s worth checking which services are covered in your chosen zone before buying into anything.
In terms of regions each provider covers, AWS has a slight lead. However, it isn’t completely conclusive So if you’re leaning towards GCP, they do still have a good range for you to consider!
Autoscaling is important for most developing businesses. It allows applications to deal with changes in demand and traffic by scaling to deal with fluctuations automatically. GCP offers autoscaling services in its managed instance groups. This helps to deal with increases in traffic but also to minimise cost when the need for resources has once again decreased.
GCP allows you to scale using various policies. These include: average CPU utilization, HTTP load balancing serving capacity (based on utilization or requests per second), and Stackdriver Monitoring metrics. You can also set a target utilization level at which to maintain your virtual machine instances. As soon as you’ve defined your autoscaling policy, you’re all set. Simple!
GCP autoscaling, like a lot of their other services, can be tested on a free trial before you buy in. But we’ll discuss pricing in a bit detail more later.
AWS also offers some great autoscaling services. Like Google’s, AWS’s autoscaling services monitor your apps and adjusts capacity when needed to keep performance steady and predictable. Like GCP, it also aims to reduce resources when they aren’t needed in order to save you as much money as possible.
AWS scaling plans are available over a variety of their resources, including: Amazon EC2 instances and Spot fleets, Amazon ECS tasks, and Amazon Aurora Replicas (to name a few). Even in AWS you’ll need to set target utilization levels for your autoscaling to work with. But, you are also able to see and manage the average utilization of all scalable resources in one place (AWS Auto Scaling). AWS Auto Scaling is free to use, and helps you minimise costs in other areas by optimising performance across your resources.
Overall, in terms of what is being offered, it’s pretty hard to choose a winner here! Both GCP and AWS offer some great autoscaling services that can help minimise costs and improve efficiency and performance on your applications. It’s a tie.
Machine Learning and AI
Machine learning and artificial intelligence are two really exciting developments that a huge number of companies are using to optimise their performance and improve the user friendliness of their services. If you’re keen to begin using it, or to develop your current uses, you’ll want a cloud provider that offers machine learning!
GCP Machine Learning
GCP offers a variety of AI and machine learning services that are suitable for everyone, whether you’ve been using it since it arrived or you’re completely new to it. These services come in three main places – the AI Hub, AI building blocks, and the AI Platform. So let’s look at what these actually offer.
Firstly, we have AI Hub. This is where you can find GCP’s plug-and-play AI components. It provides sharing capabilities that allow your business to host AI content for reuse and collaboration among internal users. Users within your organisation are able to find, use, and adapt AI components build by other teams in the same organisation. It also gives you access to AI content published by Google AI, Google Cloud AI, and Google Cloud Partners.
Secondly, AI building blocks. These are a variety of tools that help users add sight, language, conversation, and structured data to their applications. These can implement services like classifying images and improving the searchability and value of your video library.
And finally, we have AI Platform. This is a code-based data science environment. It helps machine learning developers to take their creations from ideas into realities, quickly, and at a great price.
AWS Machine Learning
But don’t think Amazon aren’t keeping up! AWS also allow you to choose from pre-trained AI services or create and use your own custom machine learning models.
Amazon SageMaker lets developers create, train, and use machine learning models quickly and easily. Or you can choose from a huge variety of pre-made algorithms and models in the AWS Marketplace for Machine Learning. AWS provides one click training and automatic model tuning. Plus SageMaker lets you train a model once, but run it on multiple hardware configurations with no issue!
On top of this, you can deploy and use your AI and machine learning services with autoscaling clusters across multiple zones. This ensures high performance and availability. AWS provides managed model A/B testing across up to 5 different models, which gives you flexibility to run experiments too.
It’s hard to say for certain, but due to the variety of services the offer, and the length of time they have been developing them it’s probably fair to give this one to AWS.
No matter what services you’ll be using, and whatever cloud provider you choose, you need one that is capable of protecting your data. So how to GCP and AWS compare for security?
Google are pretty serious when it comes to security.
They regularly undergo independent security, privacy, and compliance control verifications to ensure their service is meeting regulatory and policy objectives. Some of the many certifications they currently hold include: ISO 27001, ISO 27017, SOC 1/2/3, CSA STAR, and more.
Their infrastructure has layers of security that create a complex defence against any attacks. This basically means GCP doesn’t just rely on one technology to improve security – they rely on a number of them, so if there’s an issue with one, your data is still safe. They also offer a number of security products across GCP. The long list includes infrastructure security, network security, endpoint security, data security, identity and access management, application security, security monitoring and operations, as well as governance, risk and compliance processes.
Still feeling unsafe? GCP will automatically encrypt data at rest and in transit to minimise the risk of threats, and has an Information Security Team dedicated to maintaining their defense systems.
Google Cloud Trust Principles has also been created to help GCP’s users understand the measures they’re taking to protect your data’s privacy, in case all this has left you feeling a little lost.
So, can AWS compete with all that?
AWS also see security as a priority. Some of the services it offers are: customer-controlled encryption, connectivity options that enable private connections from your on-premises environment, and automatic encryption of traffic on AWS global and regional networks between their facilities.
They offer security support through AWS Trusted Advisors, and support with Technical Account Managers. They replicate your applications and data over multiple data centers in the same region, but you can choose to extend this and also replicate between geographic regions. You are always able to control the region your data is in.
Like GCP, AWS takes steps to ensure they are remaining compliant. Some of the certifications they hold to demonstrate this include: HDS certification, ISO 27017, SOC 1, and PCI DSS, among others!
Overall, both cloud providers do a lot to ensure your data is secure. Choosing either would mean putting your data in good hands. But in terms of the variety of security services and technologies offered, we’re going to give this one to GCP.
Ethics and Sustainability
Perhaps the ethics and sustainability of your cloud provider is a deciding factor for you. Let’s take a look at the measures AWS and GCP take to stay green.
First off, Google. Google have been carbon neutral since 2007 and, since 2017, match 100% of the energy their global operations uses with renewable energy, including their data centers and cloud. They are mainly using wind, but also solar energy. They aim to buy this renewable energy from projects that are new on the grid, to replace and reduce non-renewable energy.
Currently, Google is the biggest corporate purchaser or renewable energy, globally. They buy a surplus of renewable energy when possible in order to ensure it is consistently available. Google have reduced the non-computing energy in their data centers to 11%.
They have also been given an A rating by Greenpeace. As well as renewable energy, they have worked to make their data centers as efficient as possible, to minimise waste. They use the most energy-efficient computing network to reduce their energy use whilst maintaining the same performance levels.
They were the first company in North America to have a multi-site ISO 50001 certified energy management system. And in 2017, their ISO certification covered 12 of their data centers.
They’re working to achieve Zero Waste to Landfill in their data centers, by reducing the amount of waste generated, and increasing disposal options like recycling and reselling. Discussing every single measure Google are taking would take a long time, but if you’re keen to find out more, they publish environmental reports every year on their progress.
AWS is also committed to achieving 100% renewable energy usage for their global infrastructure. They exceeded 50% renewable energy usage in 2018, so are progressing well.
They currently have six solar farms in Virginia. Plus, wind farms in Indiana, North Carolina, and Ohio. But they are moving to expand this. Amazon have recently announced three new wind farms over Ireland, Sweden and the US that are expected to generate 670,000 MWh of renewable energy annually. On top of this, they’ve announced five new solar farms in Virginia across the counties New Kent, Buckingham, Sussex, and Powhatan.
Their newest renewable energy projects are expected to deliver 2.6 million MWh of energy annually. This energy will be able to power AWS data centers in their Ohio and Northern Virginia regions, and is the equivalent of powering 240,000 US homes in the same time period. Amazon as a whole are also making lots of environmentally conscious decisions, which they continually publish online.
So both cloud providers here are taking big steps to becoming environmentally sustainable. But, at least in part for the sheer size of AWS which got them the win on regions, the clear winner here is GCP, who have already done a huge amount to reduce their carbon footprint and their physical and digital waste.
Price is always a key consideration when choosing a cloud provider. Getting the most for your money can be a deal breaker. So how do GCP and AWS compare?
GCP has no upfront costs, you essentially pay as you go, for the services you’ll be using only. This avoids wasting money on unnecessary services. Plus, if you decide you no longer need a specific service, there are no cancellation fees!
Google automatically offers up to a 30% discount on workloads that run on Compute Engine and Cloud SQL for a significant portion of the billing month, and up to 79% off workloads that can be interrupted like data mining and processing. At GCP you pay per second of use, so you know you’re not paying for anything you don’t need at that time. You can also receive committed use discounts of up to 57%.
GCP compute engine actually averages over 8% off the price of AWS EC2, so it is a slightly cheaper option. If you’re liking the sound of GCP, but are worried about committing money when you don’t know for sure that their services are the best for your business, there’s even a solution for this! Google offer $300 free credit to spend on their services over 12 months, giving you free rein to try out the services that will be most important to you before spending your own money!
AWS is another platform that offers pay-per-use plans.
Like GCP, you only pay for what you use, and there are no additional costs or cancellation fees when you no longer need a specific service. AWS offers the option to invest in reserved capacity for certain services, like EC2 and RDS. If you choose to do this, you can save up to 75% on the equivalent on-demand capacity.
The larger your up front payment here, the greater the discount you receive. Similarly you can get volume based discounts as your usage increases. AWS supplies a Simple Monthly Calculator to help you predict just how much you will be spending before you commit.
Many of the AWS services are tiered, so the more you use, the less you pay per GB. One of these tiers is actually free and applies to over 60 of their products. This gives you the opportunity to try services in a limited capacity, to see if they would suit you, before committing your money.
Overall, both GCP and AWS offer some great pricing systems, and the option to try before you buy – which is perfect if you’re unsure about what will be best for you. GCP does take a slight lead with the variety of discounts it offers, making it a little cheaper than AWS as a whole.
For several years Gartner has put AWS ahead of GCP on both completeness of vision, and ability to execute, and our comparison here would seem to agree. However Google have been spending vast amounts of money of developing their services and it is clear that they have some specialist services that they are world leaders in so you certainly shouldn’t write them off.
This article has been pretty long, but it’s important to remember it’s still only a brief overview. If you want to find out more, all you need to do is give us a call. Because although there isn’t necessarily a “best cloud provider”, there may well be a best one for you and your business.