PIXEL PLOW. RENDER FARM SAFETY

Render farms are extremely useful when you have a deadline of a huge project and your own computer is too slow to render. It’s very convenient to be able to delegate this part of the work, and in a short time, get your final renders. As usual in the digital age, one can have concerns about the safety of such rendering services. We asked the founders of Pixel Plow render farm to help us clear the situation and address all the doubts about their render farm security.

Encryption

Proprietary software installed on the customer machine encrypts all incoming and outgoing communication using the highest encryption protocols available, which guarantees security of the data transferred. It also doesn’t depend on any third-party services that could serve as points of privacy or security failure.

Data storage

The Pixel Plow platform holds customer data only as long as is necessary to complete their render jobs. The instant the jobs complete or are cancelled, the system purges all job data. Rendered results are stored in the system only long enough to guarantee their safe transfer to customer computers, which is typically only a few seconds after each frame completes.

File transfer protocol

Since the only way to communicate with the service is via their proprietary agent software, they have implemented a unique, multi-session file transfer protocol that streams data over multiple network sessions simultaneously. This way, Pixel Plow can guarantee not only fast rendering, but also fast transfer of all job data (both for the scene file and the rendered frames) independent from the distance the customer is from their farm.

Cost security

In order to secure the final price of the renders, the system automatically suspends render jobs if the user-defined budget is exceeded. Whether submitted with a budget or not, the system displays both the Cost Consumed as well as a final Cost Estimate for each job, which gives the customers immediate feedback and control over job costs.

[ot-video type=”vimeo” url=”https://vimeo.com/117961540″]

Performance control

The slider, with 23 possible levels, of job “Power” gives users fine-grained control over the price/performance ratio of their job. The higher the Power level, the greater the number of CPU’s are exposed to their job, and the shorter the total render time. One can change the Power level up or down during the render to have “tailored” results.

Automation

The Pixel Plow platform tends to automatically handle rendering processes with no special setting or configuration required from the users. For example, use of global illumination in several render engines and applications is automatically detected. The system schedules a GI render pre-pass, storing the GI data in cache files in the job directory. A final pass version of the scene file will read in that cache, allowing the render engines to eliminate flicker in animations or stabilize lighting across a tile-rendered still image. Render job errors are also detected and handled automatically, with suspension or cancellation of the job and helpful information emailed to the user.

[ot-video type=”vimeo” url=”https://vimeo.com/153986885″]

Environmental safety

Low prices for rendering at Pixel Plow are guaranteed by the philosophy of being “hyper-efficient”. The farm is geographically located in an area with the lowest electrical power rates in the world, which also happens to be entirely sustainable, green (hydro), and locally generated by multiple sources. Their proprietary cooling system costs roughly 1/50th that of a typical datacenter both to implement and operate. Their proprietary hardware and software platform uses a minimalist approach to eliminate wasted power consumption and hardware costs. In fact, every hardware node of Pixel Plow that isn’t needed in the moment is powered off entirely automatically. This makes rendering safe both for the budget and the planet.

[ot-video type=”vimeo” url=”https://vimeo.com/154236182″]

MAIN TYPES OF RENDER FARMS

Collaborative

Different users make their computers available in order to render other people’s projects.

PRO: These farms are usually free to participate in, except for the indirect cost of electrical power.

CONS: You have to make your computer available for others, which means that you need to install render software which works even when your computer is logged off. The condition of starting your project rendering is having a certain number of credits on your account. To collect them, your computer has to have rendered other projects from users on the network. Moreover, once you send your file, it becomes public, with all the risks related. Another problem is the long queues and variable number of computers available each moment, not all with the same capacity.

Home-built

Buying several computers, one can develop his own render farm at home.

PRO: You have different computers available just for you, completely customized, and you can use them when and how you prefer, keeping your project private.

CONS: High-tech skills are required to establish everything at the beginning and the ability to fix something that doesn’t work. Initial cost of the hardware and software you have to buy, having enough space available, and the annoying noise and heat produced are some of the other shortcomings. Of course, you are still limited in compute capacity by how much you can afford to purchase and run.

Cloud/datacenter

You connect to the cloud-based farm, upload your scene, and after a certain time you receive the rendered frames.

PRO: Undoubted speed (even a hundred times more than your home computer!) because different servers are at work in parallel. Reliability: even if a computer shuts down or breaks, the work won’t be lost, but instead shifted to another server. The farm releases your computer from the CPU strain, so that you can use it for other purposes.

CONS: It is, of course, a paid service, linked to the time of use. Usually you pay by the Gigahertz or core hour.

Moreover

Pixel Plow has had many conversations with artists, directors, managers, and more in the 3D industry who wanted to have a render farm in-house. People often hold this position due to concerns about performance, security, and the cost of outsourcing. As Pixel Plow has already dealt with the performance, feature set, security, and privacy concerns in relation to using a public render farm service, the only debate remaining was about price. So, Pixel Plow decided to make an honest price comparison to prove they truly are cheaper to use than an internal render farm. They took into account several costs like: local power costs, local sales tax rates, hardware lifecycles, facility amortization time periods, local building costs, IT staff costs, software licensing, node count and type, and more.

Pixel Plow understands that computing these values is time-consuming, which is most likely why very few people ever even try. So they invite users to save time by making their spreadsheet available (you have to contact them to request a copy) to allow you to compare your costs directly to theirs.

Thanks for the Pixel Plow render farm for the information and the videos