Atlassian Data Center vs. Server: When is it worth purchasing Data Center products?
Atlassian doesn't just offer Server and Cloud versions of its products Confluence, Jira, Jira Service Desk and Bitbucket. It also offers Data Center deployments. Many customers ponder this and ask us when Data Center is a better option than Server. It's true, the Server and Data Center products don't really differ in terms of functionality.
There is no one answer to this question. It's heavily dependent on particular businesses, their processes, their priorities and, above all, how business critical the implemented Atlassian tools are for the company.
Cost: Data Center is more expensive
A central aspect during every evaluation is naturally the cost: Data Center is more expensive in all scenarios (with one exception), and it costs just as much for the initial licensing as for renewals. Unlike with the licensing of Server products, where license renewal costs 50% of the initial licensing price, Data Center products are based on a subscription model: the license fees stay the same year after year. So from the second year onward, the fees are even higher in comparison to Server. (You can find a complete summary of Data Center pricing here.)
The following table shows the license tiers of Jira software as an example:
|Data Center fees per year||Server initial license||Server license renewal||Additional Data Center fees in the first year||Additional Data Center fees from the second year onward|
|500 users||12,000 USD||18,000 USD||9,000 USD||-6,000 USD||3,000 USD|
|2,000 users||48,000 USD||24,000 USD||12,000 USD||24,000 USD||36,000 USD|
|10,000 users||240,000 USD||30,000 USD||15,000 USD||210,000 USD||225,000 USD|
(The example is based on a new license for Data Center. If you have pre-existing Server licenses, you can use them to offset these fees. Please contact us and we'll happily help you find out how much you can save specifically.)
Does the performance justify the additional cost?
The example above shows that Data Center is already initially significantly more expensive than the Server version (with the exception of the lower fees in the first year for 500 users). Because of the subscription model, the fees are again higher from the second year onward.
So is it even worth considering an investment in Data Center? The magic words are business critical. How deeply integrated are your Atlassian systems compared to how they should be? What business value is there to be gained from that? How intensively are your systems utilized?
Considering the cost aspect, customers should evaluate whether the advantages and opportunities offered by Data Center are substantial enough to offset the greater investment. The following questions and information will help with that.
Active-active clustering means that administrators can cluster more active nodes in a group of servers, on which the application runs. If a server is down, the others automatically take its place. So a server failure won't mean the failure of the whole system. Instead, the traffic will be shared among the remaining servers of the cluster.
Admins can share the load over the Data Center clusters by means of discretionary load balancing technology, whether hardware- or software-based. In addition, Data Center integrates industry standard technologies for database clustering and for shared file systems so that IT teams can use their preferred technologies to ensure redundancy on all levels of the application.
This raises the following question for customers:
How long are the current downtimes of your system and what losses do they cause you (by way of existing personal costs, loss of opportunity, etc.) per hour?
Performance and load sharing
Every node in a Data Center cluster increases the capacity for concurrent users. That puts IT teams in a position to ensure consistently good performance for a growing user base. In addition, Data Center offers application resilience, which means that a performance drop can be prevented even at peak demand because the application throughput gets increased.
Admins can configure the load balancer in such a way that the loads are distributed or assigned to different servers. You could allocate user traffic to two servers, but outsource traffic to other nodes through interfaces, automated tasks or re-indexing. The performance on the services to which users are redirected stays consistently high and will not be influenced by queries, etc.
This raises another question for customers:
How high is the current performance of your Atlassian tools? What business value would there be in increasing the number of operations possible per minute? (Duplication of operations per server is possible through Data Center.)
If organizations grow and consistently add new users into an Atlassian system, admins can add new nodes, thanks to Data Center, without shutting down the system. The indices and apps on the existing nodes will be automatically synchronised with the new ones.
The question this raises for customers is:
How steadily is your business growing? Do you estimate that not just in the short-term, but for the foreseeable future many more users will come to use your Atlassian system actively and intensively?
Updates without downtime
Data Center doesn't just avoid unplanned downtime, but also intentional downtimes. These are necessary for Server products when updating to new product versions. With Data Center, you don't have to deal with any more downtime, even with a version update. The application is updated server by server, so that if you have five servers, four are always online. Then the database will be ready to be updated. The nodes are always running and admins don't need to take the system offline at any point in the process.
The question that arises from this for customers is:
How much downtime for updates do you have per year and what losses does that incur? What additional expenditure would it take to have IT teams install updates on the weekends or after hours so that the fewest possible users are affected?
Distributed teams (for Bitbucket)
Collaboration in distributed teams, in a number of locations, countries or even continents is much easier with Bitbucket. Especially when copying of repositories, which can take a long time if teams based in distant countries are always having to retrieve the source code from German servers. Data Center lets you set up a smart mirror (a mirror server) wherever your team members may be.
Developers can pull the code they need straight from the mirror, and the German server is only contacted when the code is checked in. This can improve the performance and efficiency of distributed collaboration significantly.
The question this raises for customers is:
How highly dispersed is your software team and what difficulties does that cause?
We are your Atlassian licensing specialist
Once you've answered these questions, you should have an idea whether Data Center makes sense for your situation. But there are still more questions which may be relevant for a well-founded decision. We will gladly help you evaluate which license model is ideal for you, answer all of your questions about your licenses, and make sure that your migration to Data Center goes as smoothly as possible. Contact our license team!