Optimising Cloud spend
The Great Wall still survives, a technology dividing line between those weaned on SLAs and 5x9s and those emerging from the Cloud Native Design to Fail cocoon, fluttering the wings of network invisibility and ‘just code and launch’ mindset.
Whilst these two camps are miles apart they suffer one common blind spot, one pivotal flaw in the general scheme of things, namely the lack of technical accounting.
Back in time tales abound of 2Mb circuits running to abandoned facilities, serving no one, devoid of organisational responsibility yet leaving confused finance officers with hefty annual charges to pay. Happy days. Fast forward to public cloud compute, be it VM or Docker, and witness CFO’s operating with little or no visibility of what they are spending their cloud budget on.
A diligent operations person would point to inherent services like AWS Tagging, but does that scale? Equally solutions linking dynamic IP address harvesting into a cmdb with application definitions.
However, even if diligence prevails how do you optimise agile development requirements against available public compute offerings and associated cost models? Can resource brokers also play their part in reducing upfront deposits on elements like reserved instances?
Anothertrail have built a technical accounting reference model of vendors both commercial and open source, which when implemented addresses the cost concerns and provides detailed planning of computation resource and optimisation against an agile development environment.
This compliments our Cloud migration model covering IPAM, Packet level visibility, automation and service assurance.