A new kind of cloud orchestrator – powered by physics.
The cloud orchestration problem is massive and ever changing — with millions of options and interdependencies. Sync Orchestrator converts this problem into the only language capable of handling its scale and complexity: mathematics.
ways to launch this DAG to the cloud
Sync finds the best solution based on:
- Time or Cost
The problem starts with humans.
Despite all of the advancements in cloud computing, there exists a massive bottleneck when allocating compute resources to workloads: humans. Still to this day, humans are required to select low level application configurations, cloud instance types, and navigate complex cloud pricing models.
To add to this complexity, workloads are becoming more connected to other services, creating complex dependency graphs, known as directed acyclic graphs (DAGs). In multi-tenant use cases found at the enterprise level, thousands of DAGs are processed daily – leading to a massive scheduling bottleneck.
It is this precise combination of human error in resource provisioning and large volumes of complex interdependent tasks which leads to an enormous amount of compute waste.
The solution is math.
Orchestrator tackles this problem head on by combining Gradient with a new accelerated scheduling algorithm to provide our customers with the mathematically best way to utilize the cloud.
Orchestrator removes all manual guesswork from developers, so they can just focus on coding
Orchestrator provides the mathematical best price and performance – unachievable with current methods
Orchestrator allows businesses to dynamically change infrastructure via high level business goals such as cost and time.
To solve such a large and complex infrastructure problem in milliseconds, we borrow from a system that does this with ease – nature.
Cannonballs trace out parabolic trajectories; light always finds the quickest route between two points. The universe will always seek out the path of least resistance.
By utilizing the massively parallel algorithms employed by nature we can finally solve the cloud infrastructure problems on time scales practical to everyday cloud users. The end result? A whole new way to utilize the cloud, making it faster, cheaper, and easier for all.
Interested to learn more?
We’re building out an early prototype now. If you use the applications below and want to learn more, please reach out:
|Applications||Apache Spark, Databricks, Airflow|
|Volume Of Jobs Per Day||>1000|
|Resource Manager||YARN (Coming Soon: Kubernetes)|
|Cloud Provider||AWS (Coming Soon: GCP, Azure)|