29 lines
1.1 KiB
Markdown
29 lines
1.1 KiB
Markdown
---
|
|
layout: "intro"
|
|
page_title: "Nomad vs. HTCondor"
|
|
sidebar_current: "vs-other-htcondor"
|
|
description: |-
|
|
Comparison between Nomad and HTCondor
|
|
---
|
|
|
|
# Nomad vs. HTCondor
|
|
|
|
HTCondor is a batch queuing system that is traditionally deployed in
|
|
grid computing environments. These environments have a fixed set of
|
|
resources, and large batch jobs that consume the entire cluster or
|
|
large portions. HTCondor is used to manage queuing, dispatching and
|
|
execution of these workloads.
|
|
|
|
HTCondor is not designed for services or long lived applications.
|
|
Due to the batch nature of workloads on HTCondor, it does not prioritize
|
|
high availability and is operationally complex to setup. It does support
|
|
federation in the form of "flocking" allowing batch workloads to
|
|
be run on alternate clusters if they would otherwise be forced to wait.
|
|
|
|
Nomad is focused on both long-lived services and batch workloads, and
|
|
is designed to be a platform for running large scale applications instead
|
|
of just managing a queue of batch work. Nomad supports a broader range
|
|
of workloads, is designed for high availability, supports much
|
|
richer constraint enforcement and bin packing logic.
|
|
|