In our last update, we talked about adding CPU horsepower, changes to our storage controllers in new builds, and a firmware update. Today, we’re looking at ongoing network upgrades as part of our effort to bring all of our sites under one network design under Akamai. Our cloud computing services are historically known for usability and simplicity, but as we continue to expand capacity, we’re going to need a bigger network.
Here’s a snapshot of recent improvements we’ve made:
Global Backbone – Most of our existing sites are now connected to the Akamai Global Network, which lets technical users build distributed applications on our platform. This includes our core sites in London, Frankfurt, Newark, Atlanta, Fremont, Sydney, and Tokyo data centers with Dallas and Toronto next on the roadmap.
What does it mean for our users to be on Akamai’s global backbone? In practical terms, we’ve more than doubled our traffic delivery capacity and added extensive access to end user networks.
VLANs are now available in our Newark data center. VLANs are a free add-on that provides layer 2 network isolation for secure communication between VMs in the same data center.
Multi-Queue NIC limits have been increased to scale with the number of vCPUs up 64 (with a previous limit of 8). Multi-queue NICs enable network interrupts to be balanced across multiple vCPUs, which can potentially enhance network performance significantly.
NodeBalancers now support 10K max connections on a single port (up from 10K spread across 10 separate ports).
To Be Continued
We have a lot more planned as we keep moving forward, but these are some of the improvements you can take advantage of right now, built and ready to go.