Cloud Egress Charges: How to Prevent These Creeping Costs
One of the advantages of using the cloud is the ability to scale rapidly. On-demand scalability can eliminate the need to overbuy capacity that is only required for peak times.
However, without proper planning and oversight, cloud expenses can spiral out of control. Engineers at Adobe found that out the hard way when a single computing job on Microsoft Azure started racking up charges of $80,000 per day in 2018. A week later, they had accumulated a bill for more than half a million dollars. Capital One was also surprised by a 73% increase in their invoice from Amazon Web Services (AWS).
Surprise cloud bills can come from using more expensive compute instances, not turning off unused resources, overprovisioning storage performance and capacity for your requirements, and data egress fees. Let’s take a deeper look at the last charge — data egress fees —and how your business can reduce or eliminate them.
What Is Data Egress?
Data egress is when data leaves a network and goes to an external location. If you’re using the cloud, data egress occurs whenever your applications write data out to your network or whenever you repatriate data back to your on-premises environment. While cloud providers usually do not charge to transfer data into the cloud (“ingress”), they do charge for data egress in most situations. Another charge related to data egress fees are data transfer fees, which may be assessed when moving data between regions or availability zones within the same cloud provider.
Data egress fees are often described as hidden fees because they are billed in arrears, which means your applications, workloads, and users may continue to extract data and unknowingly run up fees before you realize just how expensive it is. It can be a challenge for organizations to monitor and manage data egress fees especially in large enterprises where multiple offices, departments, or branches all conduct data analytics.
James Orme explained it this way in Techerati: “The main public cloud providers, AWS, Microsoft, and Google, allow companies to move data into their clouds for free but charge them for moving data out again. They also charge customers when they move data to new geographical locations to accelerate speeds for their local users.”
Such charges to move large amounts of data have led to large bills. Apple spent $50 million in data transfer charges in one year. Pinterest spent more than $20 million while Netflix and Airbnb spent more than $15 million. Adobe, Snap, and Salesforce spent more than $7 million.
The cost savings of moving from capital-intensive owned hardware and software licensing to an OpEx model is one of the catalysts driving cloud adoption. However, due to unexpected fees and poor governance, realizing cost savings has been elusive for many companies. A survey in ZDNet of 300 IT leaders revealed that 37% of organizations said incremental and unpredictable costs are a top concern.
Later in this article, we have tips and best practices to reduce, or even eliminate, data egress fees. First, however, let’s dive a little deeper into the high cost of egress fees.
Egress Fees can Hinder Innovation
According to the Flexera 2023 State of the Cloud Report, 87% of enterprises are now embracing a multi-cloud strategy. Multi-cloud and hybrid cloud strategies can boost innovation for companies but may add complexity to managing cloud costs. Organizations with multiple teams working with cloud resources may also have difficulty tracking costs and budgeting appropriately.
For example, the National Aeronautics and Space Administration (NASA) generates a massive amount of data. An internal audit expects data collection to increase 8-fold by 2026 and expand to 247 petabytes. The audit concluded that the fees from moving data from a cloud present “potential risks that scientific data may be less available” and warns NASA may need to impose limits on the amount of data egress to control costs.
Surprise data egress fees from multiple cloud providers can prevent enterprises from using the best cloud provider or, like NASA, imposing data limits in an effort to reduce billing complexity. However, a single cloud strategy can lead to other risks like vendor lock-in and missed innovation opportunities.
How are Egress Fees Calculated?
Egress fees will vary depending on the volume of data you’re moving and where it goes. The more data you move, the more expensive it will be overall but the lower the per gigabyte charge will be. Location matters, too. Transferring data between availability zones or within regions will be your lowest fee. Transferring data across different regions or continents will present the highest fees.
Fees will also vary depending on how you transfer your data. For example, costs will vary on Azure whether you use the public cloud internet (ISP network) or Microsoft’s Premium Global Network. Using AWS Direct Connect can potentially reduce egress fees. AWS Snowball is a physical storage device in your data center to send data to and from the AWS cloud, although there is a per-job fee and a fee for each day on-site beyond the first 10 days.
|Typical Data Egress Charge (per GB)
|Cost to move 10 TB, per month
|Discounted Data Egress Charge (per GB) for 100 TB
|Cost to move 100 TB, per month
|Google Cloud Platform
|Free up to 10TB
|$0.0085-$0.050 depending on geography
|$850 to $5,000
Table 1: Examples of transferring data from popular public clouds at current, typical rates.
Pricing for Egress Fees on Azure
Egress fees on Azure can range from as low as $0.01 per GB for data transfer charges between availability zones to as high as $0.175 per GB.
A typical rate would be $0.08 per GB for users transferring 10 terabytes of data a month from North America or Europe to any destination using a transit ISP network. So, 10TB would cost an average of $800 a month. At 100 TB a month, the per GB price drops to $0.07 but expands the overall bill to $7,000 a month.
Pricing for Egress Fees on AWS
Egress fees on AWS (Amazon S3) range from free for less than 1 GB/month to $0.09 per GB.
A typical rate for customers transferring 10 TB a month with Amazon Web Services would be $0.085 per GB. So, 10TB would cost $850 a month. At 100 TB a month, the per-gigabyte cost is a blended rate of approximately $0.855 which would make the monthly bill $8,555.
Pricing for Egress Fees on Google Cloud Platform
Egress fees on Google Cloud can range from free for in-network moves within regions on Google to a high of $0.23 for egress to China destinations.
A typical 10TB a month transfer averages just over $0.11 per GB or results in incremental costs of roughly $1,100 a month. A 100TB transfer a month will reduce costs to $0.08 per GB and can cost $8,000 a month.
Pricing for Egress Fees on Oracle
Egress fees on Oracle Cloud range from free to $0.05 depending on geography.
Oracle does not charge for the first 10TB of data egress in a month. A 100TB a month transfer load would be at $0.0085-$0.050 per GB, putting monthly data egress charges for users at $850 to $5,000 depending on geography.
Best Practices to Reduce or Eliminate Egress Fees
You can see how egress fees can add up quickly. So, how do you prevent egress fees and avoid high data transfer pricing?
You can prevent egress fees by limiting access to your cloud data, but this is a sure way to stifle innovation and frustrate your data scientists.
Design for Flexibility
Cloud-attached storage with low-latency and high-throughput connections allows your teams to utilize advanced data analytics, AI, and deep learning from multiple clouds at the same time without having to move your data around. . You can also scale up storage when your data grows.
When designing your infrastructure, include both the cost of storage and potential egress fees in your calculations. Consider centralized storage that makes sharing among apps and clouds easy. Centralized, cloud-attached storage can also sidestep the need to put infrequently used data in the cloud at all – reducing both storage and data egress costs.
Not All Clouds Are Alike
Data egress fees can vary considerably. Each cloud has its own egress fee structure.
Use clouds with higher egress fees for only the workloads that require the capabilities of that specific cloud. workloads that demand it. When services are comparable, always choose the less expensive option.
Be Careful with Replication
Keeping data synced when using both on-premises and cloud providers is key to timely insights. However, replicating and syncing data copies across multiple providers creates data traffic that is subject to data egress fees. De-Duplicate Your DR and Backup Data Before Sending It to the Cloud
The less amount of data that is put in the cloud, the less it will cost to get it out. Before sending infrequently used backup and disaster recovery data to the cloud, de-dupe and compress it. A local edge appliance can de-deduplicate and compress your data automatically. This can reduce data egress fees when you recover data from the cloud since there’s less data being accessed.
Set Policies for What Data Is Stored
While you certainly need to save the data you need, not every piece of data generated must be saved. With the advent of data lakes that can store raw data, it can be tempting to save everything. Organizations should set policies for what data is saved based on business use cases.
Companies should also take a look at what data needs to be in the premium active areas of the cloud versus cold storage. Data with infrequent access can be stored at a lower cost, and may not require cloud-native storage solutions.
Use a Single Copy of Your Data for Access
By using a single copy of your data in a central location — and making it available to all clouds simultaneously — can cut your storage costs by 3-4x.
Large enterprises may have sufficient buying power to negotiate lower egress fees with each of their cloud providers. For less hassle, and for those without that buying power, look for data services solution providers who have already negotiated lower or free egress from specific cloud providers.
Unlock Your Data Value without Hefty Data Egress Fees
Data is one of your most important strategic assets. Unlock the true value of your data with Faction CCVs. Rather than duplicating data and storing it in individual clouds, Faction creates one copy of the data that is accessible from your cloud resources. This simplifies management and can significantly lower both your storage and egress costs.
Faction Cloud Control Volumes (CCVs) create a centralized data solution without limiting your ability to access best-in-class cloud services from multiple providers. Data is shared over an ultra-low latency connection which nearly eliminates transfer lag time.
- You can move data from on-premises to a Faction cloud via VPN, SD-WAN, or private line.
- Faction CCVs match the right performance tiers to your workload.
- Data is accessible by multiple clouds with low latency and high throughput connectivity.
- Services and applications from all hyper-scale providers can process data at the same time.
Here’s another big benefit of using Factions CCVs. Faction offers zero egress fees to transfer data out of Microsoft Azure and Oracle Cloud. Using Faction Cloud Control Volumes means you can eliminate much of your egress fees without sacrificing your data analysis.
For more information or to get a quote, contact the data experts at Faction today.