During the past few years, Amazon Web Services (AWS) has become increasingly popular as a public cloud service provider. However, other providers such as Google Cloud are catching up with it for revenue and customers.
Google and AWS are now the two leading competitors in the space for public cloud. Here we have tried to compare them in accordance with a few significant criteria so that you can understand better where each stands and make the decision in the best interest of your organization, in case you are looking for cloud services.
Among the major differences between Google Cloud and AWS is the range of services that each provides. From this viewpoint, AWS seems to be ahead of Google. AWS offers a broader range of services both in terms of quantity and quality. It builds up a substantial number of opportunities for various kinds of needs. There are tools providing specific services for media streaming and transcoding, among others.
AWS offers a managed directory service in addition to four different NoSQL and relational databases. There is even a service for cloud-based desktops that provides Windows desktops remotely.
All the different services are integrated well with each other and are part of a comprehensive solution. If you seek all your cloud services from a single vendor, AWS offers the most complete platform.
On the other hand, Google Cloud platform offers fewer services, focused on PaaS and IaaS, through its various products. Google seems to have focused its efforts mostly on PaaS. Google App Engine has been the first launched component of the Google Cloud platform. However, there are the usual relational and non-relational databases, object storage and IaaS computing as well, besides services for Endpoints and DNS.
How much effect it has on your architecture depends mostly on what your requirements are. The services provided by the Google Cloud platform are likely to be sufficient for your most common needs. Also, it is important to note Google Cloud has several strong points that AWS lacks, even if it offers fewer services.
Low cost should not ideally be the driver of decisions for buying public cloud. But companies don’t want to incur high expenses either. In the competitive world of providers of public cloud, it can be beneficial to research and comprehend the pricing structure of cloud providers.
Sub-hourly pricing and service level agreements – Where the comparison can come to a head is in the smallest purchase unit possible. In the case of AWS, the partial instance hours you consume are billed as complete hours. On the other hand, Google charges in one-minute increments after you have crossed the minimum 10-minute mark for sub-hour billing.
In availability terms, both Google and AWS offer agreements for 99.95% service. But Google’s guarantee for availability has more teeth as Google calculates it by month, instead of per year as AWS does. Google also helps users cut costs through shared core machines.
Storage – The largest instance that Google offers on demand, the n1-standard-8-dm, has 3.54 terabytes storage, 30GB memory and 8 virtual cores. A diskless option is also available in this size for a slightly lower price.
Amazon offers an equivalent standard instance on demand, which also has 30GB memory and 8 virtual cores. But users have to purchase Elastic Block Store storage separately in increments of 1GB each, whereas Google Cloud includes storage for this type of instance in its price.
Generally, Google claims that its persistent disks support up to 10 times the industry standard at 10 terabytes per volume. But if you read the fine print, you will find that persistent storage at Google costs a similar amount as at AWS.
Network pricing – Depending on the pipe size, Amazon charges different prices for inward and outward transfer of data. There are several options, beginning with free data transfer from other AWS services into an Amazon region. It begins to charge for data transfers using elastic or public IP addresses, as well as transfer from EC2, AWS service for relational databases and ElastiCache interfaces in a different availability zone in the region in the US.
Google charges a flat fee for outward transfer of data to another zone within the same region or to another region in the US. All inward transfer of data is free of charge.
Google quotas vs. AWS free tier – Although some of Amazon’s free tier offerings are indefinite, most of its free tier is limited to a year. In Google’s case, the free quotas for the App Engine PaaS are limited to a maximum rate per day and don’t extend to the infrastructure for Google Compute Engine. However, programmers can get up to 28 instance hours a month for free through the Google App Engine platform. AWS offers Elastic Beanstalk, its PaaS, for free and users have to pay only for the underlying instances.
Customers also get several other freebies with the AWS free tier, such as 750 hours of Linux or Windows t2.micro instance usage.
Both Google and AWS also provide some free network traffic. Outward traffic to other regions and services within the networks of both cloud providers are free. Free resources and hours are available with the AWS Activate and Google’s Platform Starter Pack programs.
In a number of ways, comparing prices and services between the two is like comparing apples and oranges. That’s because AWS’ main focus is IaaS and that of Google is PaaS.
While application misuse and improper planning leads to poor cloud performance, the provider you choose for cloud service also has an important part to play. So, let’s see how Google and AWS measure up.
Cloud outages – Among the main vendors of public cloud services, Amazon EC2 had the best uptime in the past year, with a total of 2.43 hours of downtime across all regions, according to an independent, third party monitor CloudHarmony.
However, AWS was behind Google Cloud in one area. Google Cloud Storage had outages totaling 14.23 minutes, whereas Amazon S3 had total outages of 2.66 hours, according to CloudHarmony.
Performance of solid state drive instances – Independent tests for benchmarking of storage performance that CloudHarmomy performed in September and November 2014 have shown that advertised performances for instances backed by solid state drives for both vendors are not consistent with claims.
You can access more details at CloudHarmony’s website.
Big data cloud services
AWS has the platform used more widely for the technology, although Google invented MapReduce. The next frontier can be streaming data analytics and Google expects its Dataflow technology to go far beyond MapReduce.
Launched first in 2009, AWS’ Elastic MapReduce service has built a strong business. Startups and big companies alike use the service to arrive at insights from huge data sets such as mapping of human genomes and purchasing behavior of customers.
However, Google App Engine MapReduce is still experimental.
The most popular big data cloud service of Google is BigQuery. It lets users run SQL-like queries on large data sets.
Google’s Dataflow is still in the beta stage whereas Amazon offers the Kinesis service that processes real time streams of data. It remains to be seen how Dataflow will fare when it becomes generally available.
AWS has a geographical distribution far more widespread than that of Google Cloud. Amazon is spread over 11 regions and has nearly two dozen availability zones, whereas Google Cloud has only 3 regions (Asia, Europe and the US) and 3 availability zones for each region.
Although Google Cloud and AWS are both cloud service providers, their areas of focus and the services they provide are somewhat different. You have to see which services you require and then decide whether to engage with either one or even both.