What (ArcGIS Enterprise) vehicle does your business need?
What vehicle does your business need? (image source: here)
I’ve always had a passing interest in cars, along with my more detailed interest in geography, maps and IT. I’ve owned several vehicles over the years but never quite managed to get ‘just’ the right one. When I’m onboarding my new customers, they are actually in the same boat with their ArcGIS Enterprise. For them, they never seem to have got it quite right. It’s at this point that customers ask Scott (geoworx) to jump in and help. It’s this place that I find my consulting sweet spot. Over the years of working with ArcGIS Server (before Enterprise), I’ve seen what does and does not work. I’ve worked with some great minds in the field, and I have been given direct access into the Esri Redlands factory (at times) to get a greater understanding.
I love the challenge of helping to take customers from zero to hero.
The other day I picked up on a LinkedIn post from a user in NZ who was saying that ArcGIS Enterprise shouldn’t be allowed out in the wild! Unfortunately, it is a common line of thinking, and I understand the viewpoint. Just because ArcGIS is not working the way you expect, doesn’t mean it can’t work the way you expect! geoworx can help you get your vehicle(s) moving.
So let’s bring this back to cars. Imagine you’re starting a new business up, and mobility is critical. What sort of vehicle do you buy? Of course, it depends upon the type of business. Let’s say you’re a fencing contractor. Around town, a van makes sense. It allows you to carry all your tools and all your materials securely, and you can bring two people to site.
But what if you’re working on a hill-farm? The van doesn't have four-wheel drive, and two people aren't enough. Do you go for single or double cab ute? A double cab sounds excellent! You’re now able to transport five people to the site, and you can get almost anywhere. However, you’ve compromised your carrying capacity. How do you get enough materials to the site? And how do you securely store your tools?
Sales are essential. You need to set the right impression. A muddy 4x4 or a van do not entirely create the right image. A branded 'run around' may be needed.
You have the capital to buy one vehicle, but you can already see that you need three. What do you do? It appears to me that the standard approach would be to get the biggest Ford Ranger you can afford! Then get some roof bars and maybe a trailer. It’s a pragmatic and sensible approach. It’s one vehicle to purchase, service and support. It just works. Perhaps, we should say – it, just about, works. The fencing gang now have the capacity, but it’s hard lifting fence posts onto the roof bars, and off again! The trailer may get bogged down on the hill. But it works, maybe, most of the time, but frankly it’s hard.
Let’s explore this above statement as it relates to ArcGIS Enterprise.
When people say ArcGIS Enterprise doesn't work, the experience in me paints a picture like the above story. As a customer, you may have used ArcGIS Server for years, and you saw that it worked well on one machine. So why is ArcGIS Enterprise so different? Well for a start, you’re installing several software components, not just one:
An Enterprise Portal, with a black-box database
A Relational Data Store including another, black-box database
A tile-cache Data Store with heavy disk operations
Each of those components has different demands on computing resources. They could need disk, CPU or memory - or a mix of all three. The ArcGIS Server is a bit of a 4x4, expected to go anywhere and do anything. Portal is the sales car, expected to be nimble, quick and look beautiful. Our data stores are the van, containing all the good things we need to do our job. So why don’t they perform like a Ferrari? I haven't seen many Ferrari's that can climb hills carrying fencing posts!
The reason that most organisations are in this situation is that they are, like the new fencing business, having to balance everything to a budget. I respect this, as I have sat on the client-side, both in the UK and NZ. I have worked in Local and Central government. I know how hard it is to get a budget and management buy-in! That’s why geoworx brings a rounded approach. When you come to me with a problem, I don’t just look for a technical solution. I try my hardest to work with your constraints to deliver the best you can afford. But 'your best' doesn’t mean you will ever get a Ferrari from a Ford Focus budget. My principal aim is to educate you and your team as to where to make any compromises. In this way, you will have realistic expectations of the outcome.
Putting all those software components on a single machine does work in some situations. But, in most cases, we need a few more servers. I have a way of explaining why this approach is needed. I do my best to avoid tech-talk, and I use picture-language and stories to reinforce the explanation. I make it understandable, to GIS, IT and business stakeholders.
I described the ArcGIS software component types above in term of vehicle types, but the most flexible component is ArcGIS Server. You can deploy it in one of several roles. That’s right, from one piece of software you can use it in multiple ways. Each role represents a different vehicle:
The hosting server should be like a Ferrari,
A mapping server should be like a BMW 330, fast as, but not quite a Ferrari.
A geoprocessing server will often take time to work and needs a grunty 4x4.
The image server can be a Ford Focus or a truck, depending upon the chosen mode, but not at the same time!
A GeoEvent server needs a high power-to-weight ratio and to be incredibly nimble, like a Pagani Zonda
A GeoAnalytics server is also a heavy-duty processing machine. It needs to be both powerful and fast.
A Notebook server can handle big and bulky tasks. It uses a different technology though, so it’s like having a Toyota in a garage of Ford’s!
Each server role looks and behaves differently. There is not a single machine that can deliver everything for you. This argument flies in the face of customers who request ArcGIS Enterprise to be cheap, easy and fast.
I believe that a coaching session is critical. I regularly ask my customers to bring their pain points to a meeting room with a whiteboard, and we break it down. In these coachings sessions, we get to the bottom of why a poorly architected deployment cannot be what you originally expected. At that point, we can choose to accept a compromise, or we start to put a strategy in place for how to deliver against your original expectation. I aim the coaching sessions just above the Esri training courses. In a Formula 1 context, it’s where the race-engineer communicates with you in the race seat to optimise the car while you’re still driving it. The driver knows how the car feels and the engineer can see the telemetry. But when the two work together, then the car (i.e. your ArcGIS) comes alive!
Once you've chosen your car, some consultants will give you a single garage. If it's what your budget dictates, then I'll do the same. Only I'll make sure there is room to expand and the initial wiring of the garage will allow you to add more garages and vehicles as you grow. I often encounter 'standard' garages with basic wiring. They are not designed to grow. Reconstruction and removals are expensive and so I work to avoid that. Some ex-army people I know have given me the mantra "Prior Planning Prevents Poor Performance".
ArcGIS Enterprise is a great tool, and I advocate it should be allowed out in the wild! The wrong car in the wrong hands is potentially dangerous though. That's where geoworx is different, it exists with a primary focus on getting the foundations right, to let an 'enterprise garage' grow and adapt over time. It's not just about having a good designer, you need someone to build it properly and then to carry that forward to be your race-engineer.
Are you race-ready? Are you getting what you need from you ArcGIS Enterprise? geoworx is ready and able to get you on the road.
Get performance-ready, geoworx helping release you on the right road (source: here)