Friday, March 11, 2011

Google cloud meeting thoughts

A few of us spent an afternoon down at Google’s Cambridge office. The topic was “100% web” and covered the Google vision of computing. I’ve been to some of these events with Microsoft both in Redmond and Waltham and have to admit, Google has a much better roadmap.

When we talk to Microsoft about the cloud and what it means the conversation turns to “opex vs capes”. While that may be interesting to some people, we are pretty clear that what we really want is faster innovation. When we told that to the Microsoft guy the response was “No, what you really want is a better QA process”, Uhm  no not really...

Will we ever realize Google’s vision of everything in the cloud? I don’t know, but at least it’s a plan that makes real sense to me. Why?

Everything is connected in the cloud.
I’d argue we aren’t there yet but everything is connectable is fair. I’m a big fan of “Innovation through Integration”. Interestingly enough I started saying this when I went to Redmond to see the “Home of the future”. Most of the really interesting stuff was integration of existing systems.

A few demos that Jeff Harris from Google showed really helped hit home with what can be done. The first one was in docs. As Jeff started typing “George Washington was born in” it automatically brought up 1732. Since docs and search are integrated, it can automatically suggest the right information based on what you already typed. Cool right? (I’m assuming 1732 is the right year, contrary to popular opinion I wasn’t there then)

The next demo was a mail merge. Jeff showed an email integrated with Google sets. This allowed him to drag down a column of cells and it figured out what the next answer would be. In his example it was adjectives to describe something. I think he had “witty, smart, clever” and google docs knew that “astute, intelligent” were also part of that set. He was also able to integrate google maps and could generate a unique map from the address for that user to get to the event.

Chrome OS rocks but can they really keep it to 8.5 second boot time? Google claims that this will only get faster. Not sure if that’s true, but since it doesn’t do anything except get me to the cloud it seems unlikely that I will need to re-install once a year to keep it running well like I do with my Windows PC. Plus it is always up to date and has no data. If I lose it, or break it, I don’t care.

Innovation can’t be installed or Speed kills
Amit talked about how one car company (local-motors?) can crowdsource a new concept car in 18 months. GM takes 6 years.  Collaboration is key to doing this and google docs helps with this with the ability to share a doc in real time.

One of the reasons innovation is so much faster with “real cloud” apps is the massive scaling and the fact that there is only one copy of the application that just happens to server millions of users. But the side effect is no customizations. You need to accept this if you want to be in the cloud. Of course customizations to SAP are what caused many companies to have problems 15 years ago, so not allowing that is probably a good thing. Plus since you are forced to upgrade you don’t get to delay it.

New ways to work are coming
Since multiple people can work on a document in real time you don’t need to make your changes, email them around, merge them in etc. Plus many web conferences are to review a document (or presentation) and that really gets merged into just regular working with a document.

HTML5 really will (should) help with web apps. I’m by no means an HTML5 expert, in fact the last time I did anything with HTML was when CSS was new and cool. But it’s supposed to let applications be more portable than they are now. VMware view, and Citrix can be written to HTML and work with any platform. This will definitely make innovation happen faster

Gadgets let you tie back into internal systems
Since not everything will be cloud for a number of years yet, gadget let you link legacy systems into gmail. We should investigate how these are built and start figuring out what users want to have “gadgetized”

Google apps are super scalable and redundant (and Geo load balanced?)
Build it once in google apps and it can scale to millions of users. Think of what that would cost if you had to build your own infrastructure for that.

IT roles may change
We really need to change from an “operations-driven” to a knowledge sharing IT org Imagine if 30% of operations staff could work with new cool technology and show people how to use it. 100% web would let us do that. The challenge with all of this is that there is sooo much new stuff it’s impossible to keep up.  If though we never had to build a new server, update an image, worry about DR, ediscovery, hardware repairs etc, we could and have cooler stuff to work on…

No comments:

Post a Comment