Joyent vs. Amazon Web Services
Downtime and latency are revenue killers. Having zero visibility into how and when outages happen is worse. Why expose your mission-critical web applications to these risks? Stop building fixes around cloud services that are inherently unstable. Consider an alternative that is designed for resiliency and stability at its core.
At Joyent, we provide a stack that reliably delivers consistent performance across a wide range of loads and can handle the unpredictable usage patterns of today’s demanding user-intensive apps. We provide a cloud-native architecture that utilizes software-driven techniques to manage compute, storage and networking resources in secure zones. There are fewer points of failure within our design, and persistent storage is not dependent on the network in order to operate.
Joyent Delivers Greater Stability and Performance
Joyent Cloud: Faults are Isolated
Joyent Cloud: Failures Don't Spread
- Zones isolate each application environment
- ZFS manages a storage pool of virtualized physical drives, if a physical drive fails ZFS automatically replicates without disruption to the compute node
- Thousands of zones can be created on one kernel
- Zones virtualize the hardware resources
Everyone's components can fail, but in Joyent's architecture, a failing component only affects the infrastructure that is directly dependent upon that component. For example, if a disk fails, ZFS automatically begins resilvering—with minimal disruption and no downtime to the compute node. If an entire compute node fails, only the virtual machines provisioned on that node fail; other nodes are unaffected. And in the unlikely event of broader failure the affected components stop, but the failure doesn't spread.
AWS: Failures Can Debilitate the Entire Cloud
- Multiple instances can be hosted on individual machines
- Hypervisor virtualizes the hardware resources
- Amazon Elastic Compute (EC2) virtualizes hardware to host individual applications
- EC2 and EBS connected via network
- AWS – Elastic Block Store for persistent storage is abstracted from EC2. It must reside in the same availability zone
- Mirroring for scaling and replication – connected via network
In Amazon's EBS architecture, a single failure can leap the fire line that traditionally separates systems: a networking fault or single component failure can debilitate scores of otherwise-unrelated EBS-attached instances. And because many have designed their systems to dynamically provision EBS volumes, EBS provisioning can find itself squarely in the recovery path: if an outage is itself EBS-induced, the time to recovery can be greatly exacerbated by the increase in EBS provisioning load.
Joyent Delivers More Computing Power and CPU Capacity per Dollar
Joyent delivers more computing power and CPU capacity per dollar than any other vendor. We don't force customers to buy more images or machines every time demand or system load spikes. And we don't force customers to compromise cost for performance consistency.
Multiplier Effect: Save More, Gain Better Performance
Joyent Cloud is the only cloud in the industry that provides you the ability to burst CPU capacity automatically. This lessens pressure for you to over-provision machines to accommodate unpredictable spikes in demand, resulting in a better economic profile for your cloud infrastructure expense.