Your data is valuable. Your organization depends on it. You worked hard to get it. You don’t want to lose it. Travis is hunched over in the corner right now cradling a flash drive and whispering “my preciousss”. But not everybody has read that book, so let’s say that your data is like your money. Everybody knows about money.
Keeping control of your data
Putting your data in ArcGIS Online is like putting your money in the bank. Both are highly trusted organizations that have developed a lot of safeguards to keep your precious things safe. Both require you to give up some amount of control over how your valuables are handled. For example, neither of them will let you run SQL queries directly against their databases.
For many use cases, that trade-off of leveraging a bank is worth it. But not always. Sometimes you need to carry cash. ArcGIS Enterprise is like carrying cash in that it gives you total control over your data in a way that ArcGIS Online cannot.
With ArcGIS Enterprise, you have options for database technologies. You decide which country's laws will govern your data. You determine how much RAM to allocate to your database machines. You get to do the updates and patch security vulnerabilities and test the backup and recovery strategy. You get to be the one who answers grumpy emails when the database is down.
You might be thinking:
“Yikes Travis, that sounds like a lot of work.”
It can be, so you don’t want to commit to the responsibility of handling your own data for frivolous reasons. You want to commit to that responsibility because it is vital to your organization’s needs.
There are several good reasons you might need that control:
- You need to integrate other systems directly with the database where your geospatial data is stored.
- You need geodatabase behaviors that ArcGIS Online data storage doesn’t provide, such as versioning or topology.
- You need to meet regulatory compliance requirements that do not allow you to put your data under the control of a third party.
- You have data sovereignty needs that require the data to remain in a specific country.
- You need to access the data on a local network disconnected from the internet.
“Travis, can you think of any other good reasons off the top of your head? No? Me neither.”
If you want to understand how ArcGIS Enterprise uses data, you should know something about the types of data stores, or containers for data storage, it uses and how where you should put them.
Categories of data stores
ArcGIS Enterprise can make use of many different data stores, so that it can support a variety of business needs. Just like you might use a coin in a vending machine and a $20 bill in a birthday card. The variety of data stores you can use in ArcGIS Enterprise come in two major categories: user data stores and system data stores.
User data stores are a “bring your own data” model for ArcGIS Enterprise. There are a bunch of supported database systems, cloud storage formats, and file types that ArcGIS Enterprise can read from and make accessible to users. Wherever your data is right now, ArcGIS Enterprise can probably read it.
User data stores are great because they give you lots of options. But lots of options for data stores make it hard for the software to be optimized for any one of them.
System data stores hold data in special formats designed to meet the specific requirements of some capabilities that really benefit from having a dedicated data store. While you control the infrastructure where these data stores are deployed, Esri chooses the underlying technologies they use and what kind of capabilities they support.
Deploying data stores
You might ask:
“I know what kind of data stores I need, Travis, but where do I put them?”
Good question!
Just like your cash is easiest to use when it’s in your pocket, your data is easiest to use when it’s closest to everything that needs it. Put all your data in the exact same place as the rest of ArcGIS Enterprise. If it’s in the cloud, put the data in the same cloud region. If it’s on-premises, put your data on-premises. If you mix the two, expect poor performance due to latency. Remember, "just because you can do something, it doesn't mean you should do it."
If ArcGIS Enterprise is in the cloud, you have the option of using managed cloud services for some of your data stores. Cloud services reduce both the burden of managing your data infrastructure and the level of control you have over that data. They can be a good middle ground between putting your data in ArcGIS Online and fully managing all your data infrastructure.
Ultimately, this whole blog post is about the trade-off between responsibility and control. There’s no way to offload any responsibility for your data to somebody else without also giving up some degree of control. Just like when you put your money in the bank. Decide the level of control you need to have and choose the data storage options that meet those needs best.
So, Travis, to summarize, we might say ArcGIS Enterprise gives you:
- the control of your data
- the responsibility of managing your data
- and options for both to help you meet the business needs your data are required to support.
Yeah, Travis, I think that get’s it.
