Organisations have turned to cloud optimisation to retain application performance and keep cloud costs under control. But despite the abundance of tools on the market, most are unintelligent services that don’t get to the root of the problem. Frank Wang from Densify, explains the problem with first-generation cloud optimisation, why many firms are failing to see results, and how one line of code can optimise all of the world’s clouds.

As cloud adoption continues to skyrocket and root organisations’ operations, cloud operation managers are in increasing need of cloud optimisation tools that match their workload demands with the right public cloud products.

Doing this manually is next to impossible. On the demand side, it’s not easy to identify demand patterns for each workload, especially over any period of time. Workloads are vast, complex, and varied. At any given time there are multiple workloads doing all different kinds of jobs, with each workload’s metrics – such as CPU, memory, and I/O – fluctuating constantly.

Organisations are struggling to see results because the tools available are immature and do not target the root causes of the problem

On the supply side, there are many capable cloud providers, such as AWS, Azure, and Google Cloud, each offering millions of possible combinations of services and service options.

Approaching this manually wastes time, introduces operational risks, and drives up cloud costs. It is these realities that make cloud optimisation an imperative for modern-day cloud infrastructure management.

Next generation cloud optimisation

Cloud optimisation is not a new concept for most organisations. Most have started to look into it and apply it to retain their application performance and keep their cloud costs under control. Nevertheless, organisations are struggling to see results because the tools available are immature and do not target the root causes of the problem.

No solution will provide good answers unless it deploys machine learning

These first-generation tools offer cost visibility: showing total spend, which department is spending what, and enable chargeback to lines of business.

Sure, a stratified cloud bill is helpful, but showing the problem is not enough. Businesses need the tools to fix it as well. The next generation of cloud optimisation uses deep permutation analysis to find the right answer to the problem of supply and demand.

Solving this problem is true cloud resource optimisation. But as it is still a developing area, organisations are not familiar with the essential ingredients.

No solution will provide good answers unless it deploys machine learning. The demand-side problem requires the creation of predictive application demand patterns. The supply-side problem can only be answered with the creation of normalised models of cloud supply using benchmarks. Bringing it all together, deep multi-dimensional permutation analysis then needs to come in and balance the two sides.

The final step is optimisation automation: the solution needs to provide cogent and accurate answers, which it can then automate. Densify’s newly launched Cloe Aware is currently the only tool on the market that offers all these steps built in. It is what we call true optimisation as code (OaC).

Smarter, quicker

By offering better answers to these problems Cloe Aware augments all the operations risk elimination and cost efficiency benefits of ordinary cloud optimisation. But with instant optimisation of the cloud, firms no longer have to worry about revisiting optimisation as their application’s demands change and evolve. Indeed, continuous optimisation (CO) will soon fit snug alongside continuous integration and delivery to facilitate a new CI/CD/CO DevOps paradigm.

OaC bridges the gap between app developers, CloudOps and finance. Developers can embed code snippets in their apps to make them self-aware and self-optimising, freeing them from the burden of determining granular resource requirements.

Cloud operations teams can provide API-based optimisation services to app teams, enabling them to run a safe and efficient environment without having to go through cumbersome processes to perform optimisation.

And, of course, finance benefits from the dramatically-improved cost efficiency that comes with machine learning, and becomes an integral part of a fiscally-aware DevOps.

By matching this explainability with a streamlined approval flow, users can make a quick and informed decision to apply or decline recommendations

Clear and transparent

Any optimisation service that builds out deep analytics and automation must not fall into the trap of becoming a black box to stakeholders. A black box approach is one that recommends actions without explanation.

With Densify, business owners can access an Impact Analysis and Recommendation Report where the performance and cost benefits and logic of every recommendation are transparently outlined. By matching this explainability with a streamlined approval flow, users can make a quick and informed decision to apply or decline recommendations.

Automated, continuous optimisation

Organisations have thus far lacked the tools to provide fully aware optimisation of their increasingly complex cloud infrastructure, relying on immature tools that only show the problem without fixing it. By building in deep analytics, automation, and machine learning into Cloe Aware’s DNA, Densify’s solution gives CloudOps clarity and insight to accelerate the benefits of cloud optimisation, continuously, regardless of how their application’s demands change and evolve.


For more info, visit the Cloe Aware website or download the whitepaper.