Moving your VFX pipeline from on-prem to the cloud is not a decision to be made lightly. The success of your studio’s entire production relies on a reliable and scalable technology infrastructure, into which you’ve already invested large sums of money and time. Even if you’ve already made the decision to migrate your pipeline to the cloud, the whole process can be a bit overwhelming - determining the optimal transition period, allocating the right budget, ensuring the right levels of access and security, as well as efficiently and effectively serving your teams across the US and/or globally.
We put together seven questions that studios should ask themselves when making upgrades to your current infrastructure. These considerations offer advice on the key decisions that should be made before your studio moves into the cloud as well as how you can choose the right cloud provider.
Question 1: Does your on-prem infrastructure limit your team?
First, let’s establish a baseline of needs. If your current infrastructure isn’t limiting your team, there’s really no reason to make the move immediately. The capex that is required to upgrade infrastructure often means that the natural step you’ll want to take is to squeeze even more out of your existing technology.
However, keep in mind that delaying upgrades for too long can lead to some serious technical debt. Rather than moving everything in your infrastructure over all at once, we recommend upgrading certain technology elements over an extended period of time to accelerate pipeline workflows while still maximizing the investments you’ve made in your own hardware and infrastructure.
Even if your work isn’t limited by your current pipeline, you should also consider the scenario when taking on ambitious projects with larger clients may require a burst solution (on-demand & overflow capacity) that allows you to expand beyond what your on-prem hardware can handle.
Question 2: How much of your pipeline do you intend to move to the cloud?
“We know that VFX studios need to amortize the cost of their tech infrastructure across many years—we certainly don’t want you to set that pile of cash on fire,” says Max Hjelm, Director of Sales, CoreWeave. “Something we’ve helped many studios do is gradually make the move to the cloud by working with their current on-prem technology.”
One benefit to working with specialized cloud providers such as CoreWeave is the option to transition your pipeline to the cloud in stages rather than all at once, giving you access to virtually unlimited rendering capacity. This is a great way to get started with cloud infrastructure while still utilizing most of your on-prem hardware.
If you decide to take a phased transition approach, make sure that you choose a cloud provider that can scale up or down with you. For example, they should offer you expanded capacity for rendering only when you need it (usage-based pricing), as well as virtual workstations for remote teams and freelancers (on-demand).
Question 3: How many workstations will you need on a regular basis and on a flexible basis?
A huge advantage of switching from on-prem to the cloud is the ability to quickly and easily spin-up new workstations for freelancers or temporary staff. Rather than keep excess render capability that isn’t regularly used or spare workstations—outfitted with software licenses—ready to go, cloud services let you expand or contract in minutes for minimal cost.
Beyond just saving on pipeline capacity or workstations for freelance artists, you’re also saving massive amounts of time for your IT team, who can get those folks onboarded in just minutes rather than days.
Question 4: Where are your teams located?
The coronavirus pandemic forced companies to shift to a remote force and even as things return to normal, many are choosing to remain totally or partially remote. If your team is spread across the globe, moving to a cloud infrastructure is a no-brainer. However, it’s important to ask where your team members are located and what their needs are. Do they need super low-latency remote desktops? If so—it would be best to choose a cloud provider with nearby data centers that can provide those high speeds.
Moving your infrastructure over to the cloud means you can also centralize your pipeline for your team across the country - improving team collaboration, transparency and pipeline efficiency. If your studio has a presence on the west coast and east coast, you no longer have to manage two entirely different pipelines in each location. Your overhead savings in both capex and management time will be significant.
Question 5: What type of security do your clients require?
If you’re currently working with larger clients or sensitive content, you’ll likely want to develop more robust security requirements and certifications.
While cloud providers themselves do not hold certifications, their security protocols will need to be up to standards if your studio is going to maintain that certification. “We work closely with studios to make sure they're able to comply with industry standard best practices on CoreWeave Cloud,” says Max Hjelm. “When you trust CoreWeave with your pipeline, you can be sure that your team has what they need to maintain your certifications."
Question 6: Are you currently outsourcing to a render farm?
If you’re already outsourcing to a render farm, you can count that cost towards your move to a cloud provider. High-quality cloud providers should be able to scale on demand as you need, eliminating the need for your typical render farm costs.
It’s also important to compare the compute costs of your preferred render farm with the cloud provider you’re evaluating. This should be close to an apples-to-apples comparison and should help you evaluate the true cost of your move to the cloud.
Besides price, investigate how they integrate with your existing storage, if at all. While some render farms may offer cheap prices and fast renders, workflows can become untenable for studios operating on a large scale without a seamless pipeline integration. If an artist must transfer large amounts of data—project files, textures, and other asset dependencies—every time they submit a job to the render farm, then it's possible they'll run into some roadblocks or delays.
Providers like CoreWeave are able to directly integrate with your on-prem storage, so render jobs can start right away. Artists can seamlessly submit a job to CoreWeave and it’s ready to go immediately—no need to wait on data transfers. This gives your pipeline a massive bump in efficiency and saves headaches for artists.
Question 7: How does the cloud provider price rendering and other services?
No one likes getting stuck with an unexpected bill, and unfortunately many cloud providers aren’t super clear with their pricing models.
When evaluating providers, look into their pricing models. Is it clear and easy to understand? Or are there upcharges for special services? Pay close attention to bandwidth costs—many providers may offer competitive pricing for compute but charge through the nose for the bandwidth required to transfer data into & out of the cloud. Another pricing consideration is how fast your provider is able to spin up instances on your behalf. If spin-up times are lengthy, you could end up paying for idle compute time that is not serving you.
With the industry’s largest selection of GPUs and a pricing model that encourages scale, working with CoreWeave empowers you to find the best performance-adjusted cost for your specific workflows, making your entire production pipeline more efficient.
Ready to learn more?
After answering these questions, you’re ready to start searching for a trusted cloud provider. With one of the industry’s widest selection of GPUs, transparent pricing, cutting edge data centers and more, CoreWeave is ready to help you move your pipeline to the cloud in a way that works for you. Contact us below to request a consultation.