Thanks to our Supporting Member Pegasus One for this thought leadership article outlining best practices for startup software products:
Optimizing cloud spend for efficiency at scale
According to a Gartner report, up to 70% of cloud costs are wasted, and a 2021 Arlington Research survey found that 82% of IT and cloud decision-makers have encountered unnecessary cloud costs—and 86% don’t feel they can get a comprehensive view of all their cloud spending.
A client approached us recently to ask our help with adapting to a spike in use and its effect on their cloud spend. Their current platform was not designed to support heavy usage, so when traffic surpassed their expectations, they elected to migrate to a cloud platform better capable of handling heavier loads. Because of its extensive offerings and reputation for high uptime, they chose AWS.
Four best practices for cloud cost optimization:
Our experience with cloud services has shown us time and time again that cloud budgets can take a serious beating if several key elements aren’t reined in. These include mismanaged or unattached resources, waste due to idle resources, inefficient monitoring of demand, and computing service inefficiencies. The four best practices we implement in cases like these are:
1. Identify unused or unattached resources – Often an administrator or developer might utilize a temporary server and later forget to shut it down when the job is over, or an administrator may neglect to remove storage attached to terminated instances. Check for open resources that can be shut off.
2. Consolidate idle resources – If an idle computing instance has a CPU utilization level 5% or less, the company might receive a bill for 100% of that computing instance—that’s a big waste of budget. To combat this, identify these instances and consolidate them.
3. Use heat maps to monitor and respond to demand spikes – A heat map identifies peaks and valleys in computing demand and uncovers opportunities for setting start and stop times to reduce costs. Take advantage of automation options to trigger starts and stops and conserve budget.
4. Right-size computing services – Analyze computing services and adjust them to the most efficient size, and also optimize for efficiencies in memory, database, computing, graphics, storage capacity, and throughput.
The Pegasus One approach:
To help our client significantly reduce AWS cloud spend, we reviewed their infrastructure against our proven best practices and implemented the following changes:
1. Implemented strategic burstable performance instances:
We spread the client’s power consumption (offered by AWS with four vCPUs, a1.xlarge) between two t3.micro instances with two vCPUs and 1GB of RAM each, delivering enough speed and memory for the client at one-fifth of the cost of the a1.xlarge.
2. Audited outbound data transfers:
We checked for increases in data transfers as usage expanded and found that a small bug in an app was triggering automated downloads that increased outbound data transfers, causing costs to multiply by 500%.
3. Eliminated dependence on elastic IP addresses:
AWS provides only five EIPs per account per region, and we alerted the client that AWS was charging additional fees each time they remapped an EIP more than 100 times in a month.
4. Deleted EBS Snapshots:
AWS provides two options for managing the state of running EC2 instances: Terminate or stop. Both options can result in lowered costs, but EBS Snapshots, which holds the information needed to create a new EBS volume with previous data. To reduce costs beyond terminating or stopping an instance, we implemented a process to delete the snapshots that were contributing to the ongoing volume to data appearing on the AWS bill.
Read this article at pegasusone.com...
Thanks for this article excerpt and its graphics to to OC Startup Council Supporting Member Pegasus One.
WHAT’S YOUR NEWS? - Submit your company news, deals, opinions, or job listings here for FREE PUBLICITY.
OR, WHAT DO YOU THINK? LEAVE A COMMENT BELOW!