The ESRI Dev Summit this year was great, as always. We much prefer this smaller more intimate conference to the much larger User Conference. They pack the schedule full of technical sessions and you have great access to ESRI developers and product owners. The warm weather in Palm Springs and the dodgeball tournament are also a nice bonus. :) Be sure to check out the free videos from the recorded sessions .
Here are a few takeaways from this year:
ArcGIS Enterprise Base Deployment FTW!
ArcGIS Server + Portal + Data Store + Web Adaptor are required to unlock all of the current and future potential of the software. Until recently it seemed that a local install of Portal was only for those with security restrictions but it is now highly recommended for the base deployment. It was recommended to install these products on their own machine since they are designed to use a lot of resources. Therefore, AGRC’s ArcGIS Server virtual machine footprint is going to expand with the Base Deployment.
The Data Store was not very well understood in our office before the dev summit. Now we know that it is a collection of databases (PostGIS and Elasticsearch) for handling uploaded portal items, raster analytic derivatives, and geoevent server items to name a few. Relational database items like CSV’s or shapefiles are stored in the Data Store in a PostGIS instance and non relational items are stored in the Elasticsearch document database. Depending on your use cases with ESRI products you can decide on which databases you need but it is best practice to split the Data Store onto multiple machines. The Data Store does not replace SDE.
We’ve been hearing about ESRI’s calcite design principles for a few years now. Calcite and it’s seems to be mature and the products using this project look really nice. The flagships are:
- Calcite Web - A generic UI framework for building web sites. Think of it as Yet Another CSS Framework (Bootstrap, Material, Foundation, Pure, Skeleton, YUI, etc).
- Calcite Bootstrap - A Bootstrap theme that implements the Calcite design principles.
- Calcite Maps - Another Bootstrap theme (and component extensions) that implements Calcite with a focus on single page applications. This project powers a new Styler AGOL template that allows app layout configuration via URL query parameters among other features.
Python Package Management with Conda
Conda is a Python package manager similar to pip with an added bonus of virtual environment management. It comes bundled with ArcGIS Pro. It’s integration with Pro and the rest of the ArcGIS ecosystem is in it’s early stages but it looks very promising. Sharing Python scripts and tools with all of their dependencies will be much easier in the future.
This is definitely the future of prerendered map services. ESRI’s implementation of the Vector Tile spec is maturing nicely and now that Pro can consume and generate vector tiles we should see them really start to gain popularity. AGRC is currently working on a version of the Overlay cache in the Vector Tile format. If you are considering switching your traditional cached services be aware that they will need to be rebuilt from the ground up in Pro. ESRI has added a handful of new symbology styles and rules in Pro to aid in the vector tile map authoring experience. Gone are the days of group layers with dozens of duplicate layers. Base map cartographers should be very happy with these updates.
The ESRI developers were tasked with creating a 100% service based architecture dataset - The Utility Network. This means that it is only accessible via a feature service but leverages a cache to improve performance. This task created ripples throughout the platform that will benefit all users. First, a new versioning model was created with the main goals of being fast and inexpensive to perform. Reconcile, post, and compress will no longer be a necessary nightly task. Attribute rules for data creation will allow for field values to be populated based on rules - Calculation, constraints, and validation rules. Feature creation templates will be stored in the geodatabase to allow for a consistent editing experience in a disconnected environment. This dataset type has potential to change and improve the way we have historically worked. Parcel Fabrics are on the timeline to also become 100% service based. I look forward to the improvements to feature services based on this work.