E&P companies utilizing cloud services for their petrotechnical workloads benefit in several ways. unmatched capacity on high-performance computing paired with increased mobility, higher information security, and improved cost control are just a few advantages. The road to the cloud might seem full of dangers, but following a set of guidelines could tremendously improve the journey.
1. Where’s your data?
E&P companies can access hundreds of data management and data science tools, a vast amount of storage technologies suited for their business needs, and infinite storage capabilities globally when moving data to the cloud. However, you need to keep your data close to the machine you log onto in the cloud to achieve the desired performance. Sessions of hundreds of MB will not run smoothly if you leave the data at the local data center and move only the apps to the cloud.
2. Assess your apps
Check that your software supports virtual machine-mode. There might also be some licensing-issues – are you permitted to run the app in the cloud?
3. Hardware requirements
One of the great benefits of a petrotechnical cloud environment is access to high-performing computing. These on-demand solutions may be notably cost-effective as they make stacks of super-machines surplus to office requirements.
Tailoring your solution includes a thorough assessment of your existing apps. Some of the most-used apps in our industry were made 30 years ago and require a CPU with high clock frequency and few cores. A public cloud is designed quite the opposite, leveraging plenty of cores and focusing less on clock frequency.
You might also have difficulty finding hyper-effective machines for the tasks requiring low latency and fast network in a public cloud. Not that they don’t exist, but they are designed for other purposes, such as Machine Learning. Accessing those machines means that you will have to pay for the full stack, although needing just a fraction of what the devices can deliver. It might come out economically suboptimal.
A centralized cloud solution for the petrotechnical workloads provides for increased security. Effective security governance, reliable, automatic back-up routines, and up-to-date access control replace local storage, portable disks, and hard-to-administer software procurement procedures.
Still, some are sceptical about giving up physical control over the workstation underneath their desk and hand the data over to the cloud. However, focusing on how you put data into the system and how you access them, a cloud solution can increase your security. Besides, a central system gives you full control over who accesses the different apps and their coherent data.
When the workstation is in the cloud, a standard laptop will be sufficient for remotely accessing any petrotechnical workflow. Consequently, you may work from home, on travel, or wherever you prefer.
All data and projects are accessible 24/7. You never need to go to the office at night only to check if the massive calculation or the rendering of a project is still running. Services running from central servers instead of local workstations give G&G professionals guaranteed uptime and put the responsibility on the cloud provider.
6. Test environment for new apps
Thorough testing of new versions and upgrades of apps is possible from a dedicated group of virtual machines. A test group may work on quality checks and discover compatibility issues in an environment separated from the everyday workflow. When given the green light, the apps are updated centrally and deployed to all workstations in production.
7. Cost control
Cloud is not about cost reduction but cost control. That might come as a surprise to someone. Still, a cloud solution's economic benefits include the right capacity for the different workflows, apps procurement management, and not paying for overhead capacity when not needed. The scalability of a cloud solution gives you predictable estimates for your data storage and compute consumption.
Geoscientists often face massive calculations or other processes that occupy their workstation for hours. In a cloud solution, you can log on to a high-performance machine built specifically for these kinds of operations and let it run until the process is finished. Meanwhile, you can log on to your regular day-to-day machine in the cloud and continue other operations there.
When you need high capacity and deadlines are tight, more processor power and machine capacity are available to you. It comes with a cost, of course, but this upscaling is a lot easier in a cloud solution.
9. Make your own applications vendor-agnostic
The typical setup in a company in the G&G business includes several native apps, like job servers and web front-end. Rewriting them with a general layer makes them vendor-agnostic, so they can be used on whatever platform you decide, whether it’s Google, Microsoft, Amazon, or others.
When you face a situation where you need to run several costly GPU or CPU-intensive processes, you can approach the cloud vendors and negotiate the processing power you need at the lowest spot rate.
10. Plan the workflow carefully for hybrid solutions
A hybrid solution is when you decide to move just parts of your workflows to the cloud and keep some functions in-house. Often these scenarios consist of a data set at one cloud provider while buying a most wanted service from another vendor, in combination with a few processes running at your local datacenter.
Knowing how the apps are connected, the data flow, and size is critical to a seamless integration between the platforms in a hybrid setting.
These are the most challenging scenarios in a hybrid setup. Transferring data in and out of public clouds comes with a cost. Performance may be reduced, and moving data out of a cloud might have an economic impact.
Cloud or not cloud
The CIO needs to assess limitations, differences, and cost models for a cloud compared to a traditional on-prem solution. Some services in the cloud are cheap, while others can be costly. A move to the cloud solves particular problems, but others still exist. The cloud is not a magic formula.
When everything is measurable, moving to a cloud solution is more about cost control than cost reduction. By carefully planning for peak capacity demands and only using the necessary capabilities when needed, the company can reduce investments and still get the resources required.