Amazon Elastic Compute Cloud (Amazon EC2) is a web service that provides secure, resizable compute capacity in the cloud. Users can launch instances with a variety of OSs, load them with custom application environments, manage network access permissions, and run images on multiple systems.
$0.01
per IP address with a running instance per hour on a pro rata basis
Amazon SageMaker
Score 8.2 out of 10
N/A
Amazon SageMaker enables developers and data scientists to quickly and easily build, train, and deploy machine learning models at any scale. Amazon SageMaker removes all the barriers that typically slow down developers who want to use machine learning.
N/A
Pricing
Amazon Elastic Compute Cloud (EC2)
Amazon SageMaker
Editions & Modules
Data Transfer
$0.00 - $0.09
per GB
On-Demand
$0.0042 - $6.528
per Hour
EBS-Optimized Instances
$0.005
per IP address with a running instance per hour on a pro rata basis
Carrier IP Addresses
$0.005 - $0.10
T4g Instances
$0.04
per vCPU-Hour Linux, RHEL, & SLES
T2, T3 Instances
$0.05 ($0.096)
per vCPU-Hour Linux, RHEL, & SLES (Windows)
No answers on this topic
Offerings
Pricing Offerings
Amazon Elastic Compute Cloud (EC2)
Amazon SageMaker
Free Trial
No
No
Free/Freemium Version
No
No
Premium Consulting/Integration Services
No
No
Entry-level Setup Fee
No setup fee
No setup fee
Additional Details
—
—
More Pricing Information
Community Pulse
Amazon Elastic Compute Cloud (EC2)
Amazon SageMaker
Features
Amazon Elastic Compute Cloud (EC2)
Amazon SageMaker
Infrastructure-as-a-Service (IaaS)
Comparison of Infrastructure-as-a-Service (IaaS) features of Product A and Product B
Suitable for companies that are looking for performance at a competitive price, flexibility to switch instance type even with RI, flexibility to add-on IOPS, option to lower running cost with the regular introduction of new instance type that comes with higher performance but at a lower cost.
Amazon Sagemaker suits well in areas of data science and Machine learnings where medium to high-volume data is to be used for analysis. For a lean and platform agnostic deployment, it provides kubernetes integration to containerize the solution and deploy on any platform. It is one of the best solution for technical users for training Machine Learning models.
A great variety of choices in Amazon Machine Image (AMI) types. Users can select a more basic type to run generic workloads, but also have the choice to pick an AMI pre-installed with specific services in the AWS Marketplace.
The range of instance types can support the usage from a student's exploration (inexpensive general-purpose nano instances) to an enterprise's most intense workloads (memory or storage-optimized instances with terabytes of memory and ultra-fast network connection).
The pricing options, from regular instances, reserved instances to spot instances allow users to get the job done and make smart choices about how much they want to pay and when they want to pay.
SageMaker is useful as a managed Jupyter notebook server. Using the notebook instances' IAM roles to grant access to private S3 buckets and other AWS resources is great. Using SageMaker's lifecycle scripts and AWS Secrets Manager to inject connection strings and other secrets is great.
SageMaker is good at serving models. The interface it provides is often clunky, but a managed, auto-scaling model server is powerful.
SageMaker is opinionated about versioning machine learning models and useful if you agree with its opinions.
This service is a bit difficult to consume. New users need a big learning curve to use this service effectively.
UI for EC2 service is a little complex and at many places, it misses detailed explanation.
Sometimes it takes too long to create images of EC2 instances. This keeps your EC2 up for that extra time. When instances are heavy, it penalizes a lot of money.
It's easy and straightforward for a technical person to use it via SSH, but when working in cross-functional teams, using Amazon's web console is difficult for this particular service. Most modern cloud providers provide a more seamless user interface to interact with their cloud machines, and the same should have been the case with EC2.
AWS's support is good overall. Not outstanding, but better than average. We have had very little reason to engage with AWS support but in our limited experience, the staff has been knowledgeable, timely and helpful. The only negative is actually initiating a service request can be a bit of a pain.
Azure VM Builder offers good service, but the options are quite limited (Too much inclined to Windows as it is prepared by Microsoft). EC2 image building capabilities are the best in the market, and offer Windows, Linux (CentOS, rh2, debian, ubuntu), along with other distros, which helps customers choose according to their needs.
We have not invested in another machine learning software at this time and so far this has proved very successful with our machine learning teams. As mentioned, I am training these individuals simply on the fundamentals of the software and using it/customizing it for their needs. It has been very easy to do this and has gotten great reviews across the organization so far.
With EC2 you pay only when is Running, so you can save up to 75% on Dev environments which are running only on office hours
You have several ways to pay for EC2, with EC2 Reserved Instances you pay with a discount of up to 72% if you make a commitment of using them from 1 or 3 years
With EC2 spot you can use spare AWS EC2 capacity with a discount of up to 90%, your workload must be interrupt tolerant as your EC2 could be reclaim by AWS and the EC2 terminated
