Atlassian Cloud VS Server Evaluation

[Prashant Prayaga]

Team,

I would like to have detailed comparison in features that Cloud VS Server for BAMBOO,JIRA,CONFLUENCE,BITBUCKET.

[Dave Theodore [Coyote Creek Consulting]]

What specifically are you looking to compare?

When we have this discussion with clients, here are the differences that are the most meaningful to the majority of people.

• Bamboo is not available from Atlassian in Cloud, but you can run it in AWS, etc.
• Jira and Confluence both basically behave similarly in Cloud and Server.
• Not all Apps are available for Cloud, so if you have a need for App provided functionality, it's worth perusing Marketplace to see if the ones you need are available for your platform of choice.
• Cloud has pretty significant storage space limitations. This seems to be the biggest deal breaker for most clients we talk to.
• Cloud is limited to 5000 users.
• Bitbucket is quite different in Cloud compared to Server. You can only install Apps on Server.
• Read up on Atlassian Access if you go with Cloud. This helps close the user management gaps between Server and Cloud.

If there are other things you're interested in, ask more specific questions and we'll be able to help you better.

[Prashant Prayaga]

 HI Dave,

Thanks for your reply.

As mentioned earlier , we are planning to incorporate CI/CD tools like BAMBOO,BITBUCKET, JIRA and other tools. So we would like to evaluate both the offerings from Atlassian.

Intially it will be for small group of less than 500 and if gets success we can have more than 10k users.

Your recommendation would be highly appreciated.

Thanks,

Prashant

[Nic Brough -Adaptavist-]

Get the server version.  Cloud does not support more than 5,000 people.

[Dave Theodore [Coyote Creek Consulting]]

+1

Based on our internal testing data, I feel pretty confident that you can run Confluence in AWS for 10k+ users (depends a bit on your usage patterns, installed Apps, etc,) but I wouldn't not recommend running such a large Bitbucket instance in AWS.  With all the Data Center products, you need to use shared storage to house the shared data. The latency numbers that I have seen on the current AWS and Azure shared storage options indicate that you will have significant issues running a large Bitbucket instance on them.  You are at high risk of encountering queueing issues with such high latency storage. Git generates a lot of diskio when branching and merging, so it needs really fast, low latency storage in order to run at scale. This shouldn't pose a major problem for Confluence or Jira, as they are not heavy diskio users for normal operations.  I would expect to see reindexing slowness, though.

With such a large instance, don't cut corners. Do a proof of concept, load in some data and do a lot of testing before releasing it in the wild.  Just building things and hoping for the best is the wrong approach for a deployment like this. You're talking about a $500k+ expense per year when you factor in Cloud service provide cost and licensing. Anyone in your organization with that kind of signing authority will expect things to work flawlessly with such a large yearly expense. You may want to consider working with a Solution Partner that does this sort of thing for a living and has dealt with this sort of challenge at a large scale.