Network Time Foundation uses a private cloud for its public services and internal development. We are often asked why we use our own, private cloud instead of one of the commercially available public cloud options like Microsoft Azure, Amazon Web Services, or Digital Ocean.
The answer is simple: it is more cost-effective. Let’s take a look at a comparison between what it currently costs NTF to host its own equipment versus utilizing the Amazon Public Cloud. (Note: Amazon cost estimates are from our most recent calculations.)
Cost Comparison: Amazon Public Cloud vs. Private Cloud
There are three cost components considered in calculating monthly costs: Storage, CPU Utilization, and Bandwidth.
Storage costs for Amazon S3 were calculated at $23.50/month per TB. NTF currently uses approximately 650TB of storage across five locations.
Storage cost to NTF for utilizing Amazon would be $15,275/month (650TB at $23.50/TB/month).
Storage on NTF’s private cloud costs $0/TB/month.
Ten Amazon EC2 instances would have a recurring charge of $18.70/vCPU/month. This does not include bandwidth charges. The NTF private cloud has over 600 vCPU cores and nearly 1TB of RAM.
CPU cost to NTF for running on Amazon would be $11,220/month (600 vCPUs at $18.70/vCPU/month).
CPUs on NTF’s private cloud cost $0/vCPU/month.
Amazon bandwidth costs run approximately $30/Mbps/month (megabits per second) of consistent data transfer. NTF’s private cloud provides 3000 Mbps of unrestricted bandwidth at three data centers and can obtain additional bandwidth at no extra charge. (At a fourth data center, NTF is 100% virtual using a very high-speed internal network, at no cost to NTF.)
Bandwidth costs of running on Amazon would be $90,000/month (3000Mbps at $30/Mbps/month).
Bandwidth for NTF’s private cloud costs $0/Mbps/month.
Here’s a summary table:
|Component||Amazon Cost||NTF Private Cloud Cost|
|Storage||$ 15,275/mo||$ 0/mo|
|CPU+memory||$ 11,220/mo||$ 0/mo|
|Bandwidth||$ 90,000/mo||$ 0/mo|
The decision is obvious. The most cost-effective hosting strategy is for NTF to continue to maintain private cloud resources in its own data centers.
This is one way that NTF makes the most of its funding—we spend our money on projects rather than on overhead, operational, or capital expenses.