Saturday, December 02, 2006
AGX vs. GE & VE3D: Competition?
Posted on Saturday, December 02, 2006 2:19:32 PM (Mountain Standard Time, UTC-07:00)  Comments [6] | 
Categories: ESRI
This started as a comment over on James' blog, but it got a little long for just a comment.

"gonzo" had replied to the comment that “ESRI has no network infrastructure to compete with Google.” with this:
This point could be argued, considering that ESRI services are published from AT&T data centers in Virginia and New Mexico, also offering triple redundancy of power and geographic redundancy.

[UPDATE 12/7/06: upon re-reading this post today, I concluded that the section I have removed, did not add to the conversation.]

I'd just like to point out that while ESRI hosts servers in  "AT&T data centers", Google and Microsoft have multiple data centers, with 100's of thousands of servers. While this infrastructure is not dedicated to their mapping applications (I'd hesitate to call them GIS's), they can leverage these to provide a level of service that's almost unparalled.

Thus, while I commend ESRI for hosting public globe services, I think it's unrealistic for users expect the same level of data streaming performance from these services that you will get from Google Earth or Virtual Earth.
Saturday, December 02, 2006 9:32:16 PM (Mountain Standard Time, UTC-07:00)
Wow!
A surprising response to a simple AGX post…and abit taken out of context.

I will not approach your post at the edges, rather the center. The numbers are too complex and my point was not that ESRI can out maneuver Google or Microsoft. Although, many of those previously mentioned servers could be caching the over 8 billion web pages?? In addition, Google is not Google Earth/Maps and Microsoft is not Virtual Earth/3D, so you may as well be telling me about the profit line on apples versus oranges. Right! Big Deal.
All at the edges…


So at the core, your point was that these companies don’t care about GIS and that there isn’t any “$$” to make it worth wild. Well, arguable these companies have already began their opposition. What all the fuss is for if this isn’t a struggle? To illustrate the point, note the maniacal spending spree to buy Keyhole and global imagery that covers two/thirds the world’s population or 3D models over 22 US cities or Sketch-up or the rest…Both Google and Microsoft are pursuing the consumer market share. While G & M go after consumers ESRI pursues the large and directly paying business sector. This sector now is worth about 600$million per year, but the “worthwhile” sector is yet untapped. Geospatially Enabling Technologies (GET) is what ESRI is after and it offers markets that are orders of magnitude larger then silly GIS professionals. ESRI has strategically placed themselves into position, marring many enterprise software vendors and deploying an application that fits well into SOA & ESB.
Gonzo
Sunday, December 03, 2006 5:38:51 PM (Mountain Standard Time, UTC-07:00)
Admittedly, I did get a little carried away. ;-)

My point is that if any of the really large players decide they are interested in geospatially enabling the enterprise, they can and will bowl over ESRI. Yes, ESRI has a really wide and deep toolkit, but this comes at the expense of performance and agility. You can do nearly anything with Server, but that means that it's optimized for nothing. (specifically raster processing is a dog) If Google or Microsoft sees a new "killer app" that involves spatial integration, I believe that they will do what Keyhole did - focus on that one thing, and do it really really well.

I totally agree that the bulk of the market is still untapped - Google/Microsoft/Yahoo have taken on the free-maps-with-ads thing, and ESRI/MapInfo/Autodesk have the niche "GIS Professional" market. The question is - who will figure out how to grab a share of the rest of the pie? Maybe it will be ESRI (better positioned than other GIS vendors), but only if things move slowly, and the big guys decide to stay out of the market.
Dave
Tuesday, December 05, 2006 9:36:44 AM (Mountain Standard Time, UTC-07:00)
Do not sweat the previous post, your zeal is admired.

Well, is it too diplomatic to say that I agree and disagree with your most previous post???

Your right that ESRI does many things well, and maybe lacks that focused market spot, and probably correct when saying that Google/Microsoft/and the rest are better positioned to capture any line that they desire. With all that, I think that you are wrong on all of those points too. Contradiction, yes...

This is why…those untapped market are the folks that have either inherently spatial information or information that has a small spatial component. This point is difficult to argue. This is an information rich society, and 80-90% of that information has a spatial component. Those folks use large-scale enterprise systems that are all about interoperability, reuse, standard description, and discovery. Also, componentized and fragmented applications, modular degression and the rest…or also known as Service Orientated Architecture, coupled with the Enterprise Service Bus. That is what AGS is, at an abstract architectual level, SOA. That is the point that is often over looked when discussing AGS, we look at the surface and say things like; “that looks a lot like %*^*&”. Does it matter?

AGS is their product that is focused and focused at large scale enterprise systems. Look carefully at the changes in this release and you will see many SOA piece falling into place, also many new business partner such as SAS, SAP, IBM, and so on.

Gartner documents that 80% of all applications by 2008 will be developed using SOA. AGS is there and first to market. Is this coincidence or practice by ESRI?
Gonzo
Tuesday, December 05, 2006 3:23:18 PM (Mountain Standard Time, UTC-07:00)
All good points, but I guess I'm feeling that SOA & ESB are more hype than reality - particularly when GIS is involved.

The SOA re-use and plug & play stories are great for vertical markets where standards have been developed. But GIS does not have this locked up yet, and I'd argue that we're a ways from it.

The OGC standards are, for the most part, not SOAP based (i.e. getting a map - WMS). Using GML to pass geometries around is near insane (try validating a feature against the GML schema - it takes ages). Not withstanding other issues, these items alone take two large chunks out of the picture: creating maps and passing geometries around.

So, this leaves the creation of very simple services - locate parcel by PIN, by Address, by Owner type of thing. And in most cases, the group implementing the service would have to use a custom implementation (8 digit PINs, vs 9 digit PINs etc)- which blows the plug & play out.

Which leaves the SOA/ESB story with "re-use", and studies have shown that is very rarely acheived in any sort of software development. Basically, if you can't write COM/C++/Java/.NET components which are massively re-used, how does adding WS-* to the mix change the underlying functionality? It doesn't.

So - while I'd agree that ESRI has a lead in getting to the SOA/ESB party, I think that nobody really knows what to do next! Maybe the result is that ArcGIS Server becomes the defacto standard since it is the only thing in the marketplace, but this still does not help out with the plug & play aspect if there is only one vendor!

Anyhow - all very interesting stuff, but I've got some work to do (and a post on System Design Architecture to get completed!)

Also a note on the un-tapped market: DB2 is the most widely deployed database in the world (just ahead of Oracle). Yet ESRI has very very few DB2 users. Find a complelling business case to sell GIS to those DB2 users, and you've got that un-tapped market segment!

Cheers,

Dave
Dave
Wednesday, December 06, 2006 8:02:22 AM (Mountain Standard Time, UTC-07:00)
Working for a competitor to ESRI, it's clear that Google Earth does present some potential problems for GIS vendors like ourselves. However, Google Earth has an 'Enterprise' option that looks anything but - it's still missing a lot of GIS functionality. There's, for the time being, still plenty of opportunity for existing GIS vendors to run against Google. Even if Google does increase their capabilities to have a more fully-fledged GIS system, then you'll still find it possible to compete by finding your unique selling point. Photoshop may be the biggest name in image editing, but Paint Shop Pro still managed to make a name for itself. Equally, GIS isn't a market where just one vendor is likely to own the entire market.
In the UK, ESRI and MapInfo have both lost many potential customers to small GIS vendors, particularly in the Local Authority market. These competitors are companies whose products are now common in significant proportions of some markets (such as the UK Local Government), despite the companies having revenues of <£1 million p.a. ESRI may be surprised by finding itself the smaller fish for a change, but there's plenty of room for multiple offerings in the marketplace.

What should be important to the GIS vendors like ESRI and my own employer is that while Google Earth doesn't do everything you want it to, it does do some things better than most of the competition. Their navigation, integrated with SpaceNavigator, and using reasonably smart caching and 3D acceleration, makes a mockery of the drawing times of most of the major GIS systems. Features which stand out above the competition WILL force the existing product developers to look at their offerings and adapt so that they still offer something that can readily be seen as better than Google Earth or the like.
David
Wednesday, December 06, 2006 3:58:59 PM (Mountain Standard Time, UTC-07:00)
Dave, great post.

The hype is the only real thing attached to SOA, in fact, Gartner has a "Hype Cycle" (Don’t know if you have seen it, since its not cheap) and SOA is at the summit of the "Trough of Disillusionment”.

FYI. IBM licenses ESRI technology to extend DB2’s spatial capabilities. ESRI & IBM have a strategic partnership that extends ESRI technology further into base-level IT infrastructure. This is in addition to the previously mentioned enterprise vendors.

Looking forward to reading your System Design Strategy post…

Vr,
Richard
Gonzo
Comments are closed.