When it comes to cloud, networking is the backbone. This is because users constantly need to be connected to the internet and push data over cloud platforms or on-premise hardware. The data is communicated over different regions too, and with the data moved around, there are network cost implications on the organisations. The cloud computing services like Google Cloud Platform and Amazon Web Services have always dealt with the challenges of trying to reduce the networking cost.
GCP is fast becoming the crowd favourite with offers such as committed usage discount when compared to its rivals like AWS. GCP also offers one of the top services to both beginners and experts. Networking being a key element of managing cloud-based services, it is always great to have a reliable routine that would cost and optimise the workflows.
Here are a few best practices as recommended by the experts at Google cloud:
Directing Traffic Flow
The first thing to understand when it comes to networking cost optimisation is to monitor what goes in and out of the cloud platform. A lot of emphasis is given to network traffic flow, and one of the ways to govern your network traffic is by a feature VPC Flow Logs. Now, VPC Flow Logs keeps a filing of your network flows sent and received by the VM instances. Each of the log entry record details is information needed for understanding network traffic.
These log entry details such as source IP, destination IP, and bytes sent and received for each network connection are collected in Stackdriver logging. They can be exported to BigQuery, where you can visualise the trends.
The VPC Flow Logs can be used in cost optimisation and other cases like network monitoring, forensics, real-time security analysis.
Also, while optimising your network cost regarding VPC Flow Logs, Google recommends to keep these three things in mind:
- Traffic between zones and regions.
- Top talkers.
- Traffic to specific countries on the internet.
Different Region, Different Charges
When it comes to saving on the actual network costs, one of the factors to emphasise on is the geo-location, whose information is taken from the VPC Flow Logs. The geo-location is an essential factor when it comes to network spending because not all network charges are equal, different region, different costs.
- Along with using VPC Flow Logs, one can make use of Network Intelligence Center, a network monitoring, verification and optimisation app, which allows gaining access to the network bandwidth used between regions and geo-locations.
Gaining a better understanding of one’s traffic patterns across regions while transferring data around the world is crucial, whether it is to a customer or other internal services in your cloud platform like GCP.
- While giving importance to the regions, one has to consider the zones that their workload lies in, as one can architect them with no charges while using a network between zones. By configuring public, external IP addresses to communicate via the zones’ internal IP address, the organisation can save the cost of what they would have paid that traffic.
One advice from Google cloud team would be to use a single-zone architecture in regions with higher costs and multi-zone architecture where the traffic costs are lower. This comes to light when the organisation will need cost saving with availability implications of a single-zone architecture.
The Dilemma Of Using VPN
From VPC Flow Logs, the organisation will know what kind of applications are using the most bandwidth, which means reviewing your total bytes transferred. If one is pushing large amounts of data, they may be entitled to potential discounts.
If an organisation pushes data daily and has databases dedicated on-premises hardware and frontend applications serving requests in GCP, then one can make use of a Dedicated Interconnect or Partner Interconnect.
And if they’re pushing TBs and PBs amount of data regularly, then it can be cheaper to opt for a dedicated connection compared to accruing costs associated with traffic over a public network or VPN.
Network Service Tiers
Google’s Network Service Tier offers a trade-off between performance and cost. By choosing either Standard or Premium Tier, one can potentially reduce the charges on services that tolerate more latency and don’t require a service-level agreement (SLA).
Although choosing Standard Tier may have low pricing, the organisation needs to determine whether it can benefit from the lower pricing without impacting the performance of services.
Pay For What You Use
Pay only for what you use is one of the most significant benefits of the cloud. Keeping this in mind, here are some recommendations by Google to follow and make the most of the investment pertaining to GCP users:
- Enabling and customising logs like VPC Flow Logs, Firewall Rule Logging and NAT Logging.
- Gaining private access for the enterprise to reduce cost when possible.
- As a general security best practice, it is recommended to use internal IP addresses where applicable.
Enterprise-level use of cloud means one key thing — the lower the cost, the cheaper it is to maintain. This is simply because cloud service is one of the significant cost contributing factors in an organisation. Since cloud computing is the future, and the global cloud computing market is set to reach $258 billion in 2019, one-third of an organisation’s IT budget is allocated for cloud services. Another stat which further stamps the global impact of the cloud is that 90% of the companies are on the cloud, and it seems like the entire world is moving on to the cloud.