Google has done an excellent job of making themselves an indispensable part of internet life. Most of us couldn’t imagine a doing our jobs without it anymore. But this begs the question, why don’t they act like it.

This morning, like most mornings, I did a google search for 3tera looking for a few things; who’s advertising on the keyword 3tera, are there any new entries on the first couple pages, and how many search results are shown. Why do this so frequently? Partially because it helps me track our PR effectiveness. However, the larger reason is that I believe my customers use google to validate 3tera and I want to see what google’s showing them.

You may note that I wrote validate and not research, and that’s intentional. When dealing with a new customer, partner or even competitor I do a google search and look to see what type of links show up on the first page. Are there articles in serious journals or just their web page and some blogs. Plus I look at the number of search results google presents. Hundreds of thousands implies a critical mass of interest. Just a few thousand means I’m dealing with someone nobody talks about. I know others see google as validation as well because I’ve gotten emails when an article about 3tera adds tens of thousands to the google search results.

Having done the search countless times, I pretty much know what to expect. However, this morning I was puzzled, because our search results had dropped from around 230K, to less than 70K. What had happened?

After a bit of research, I noted that several other companies I track had similar drops in the number of reported results. Evidently, google changed the way they count their index. So what? Well, it turns out this isn’t the first shift I’ve had in google data recently.

I got a little shock a couple weeks ago when I logged into Google Analytics to check our web traffic - the reported value for Time-On-Site was half what we normally run. I quickly checked our website to ensure we didn’t have some error on the server, but it was running fine so I went back to Analytics. Pages-per-visit and bounce rate were normal. Only the time-on-site was different. I looked up a couple historical figures and they too were low. Analytics was showing our historical figures at roughly half what we’d recorded before. Evidently Google changed how Analytics calculates time.

This has happened recently with other online services as well and it feels like a trend. It seems google and many other online service companies feel their applications exist in a vacume and that they should therefore be able to make whatever changes they like. The engineers and product managers at Google, for instance, probably updated the Analytics time calculation because they felt it was more accurate, but that shouldn’t be the end of the discussion.

Many companies, 3tera included, use online services like Analytics to track their business. I don’t believe that the number of search results or the time-on-site are precise measurements, but I do expect them to be consistant. That consistancy allows me to view their rise or fall meaningful.

Google and other service operators need to recognize the data being generated by their applications is used outside of their systems and they need to ensure consistancy of their results.

UPDATE 10/2/07

Well, four days later, the number of search results Google’s returning for 3tera are back to where they were last week. Most of the other searches I ran are also back to normal. While I’m glad to have the system running as expected again, I’m still curious what happened.

Layered Tech has taken the covers off their new DynaVol storage service.

In addition to offering a reliable and secure way to store data, DynaVol can also be used to serve static content, taking a lot of load off your infrastructure. Plus, with packages starting at just $15 month, the system is exceptionally affordable.

I’m looking forward to using DynaVol for near-line storage of application backups.

Canadian domain registrar and hoster Sibername.com announced on their forum today that they’re beginning implementation of AppLogic. We’ve been working with them since HostingCon in Chicago as they put AppLogic through a barrage of tests. They have a good team and we’re pleased to have them on board.

When I was in college the irreverant comic strip Doonsebury was in it’s prime. Because it featured college kids it was a favorite on campus. One frequent story line involved the young college kids workig as polsters during election years and in a particularly memorable line the polster responds to a woman who’d never heard of his candidate with “he’s never heard of you either.”

So, here I am going through that same dance with other startups at AjaxWorld. Sure, Oracle and IBM are here along with a few other industry stalwarts, but a large chunk of conversations between exhibitors is “so what do you guys do?” Ah, the life of a startup.

For startups, how you answer this question is an important decision, it’s your elevator pitch to use the venacular. What three sentences can you string together to convey what you do. If you’re building a faster, smaller, cheaper widget then your answer is pretty simple. On the other hand, if you’re building something new, as we believe 3tera has, then the job gets a little tougher.

Imagine showing a spreadsheet for the first time, a GUI, or maybe a mouse. I’m sure showing the first web browser stretched a few imaginations. Each of these is well understood today, but when first created the developers had to find a way to communicate. You search for a meme you can associate with.

I just did my video interview with sys.con.tv here at AjaxWorld and have a panel session this evening - all preparing for the day when no one will have to ask what we do.

After a couple decades building technology, you come to realize that technology development is really a cycle rather than a vector. For instance, the way in which AppLogic packages an application and its infrastructure is analogous to the way you’d launch an app on an SMP system with a shell script that starts processes and connects them through sockets. Jason Brooks has a short but interesting piece about this subject, and the way in which the iteration adds value.

Often the parallels aren’t obvious at first, because you start out trying to solve a problem rather than recreate a technology in a new space. As an example, we weren’t trying to create an operating system when we started AppLogic - we simply wanted to help folks scale web services. We wanted to give them a way to express the application structure so it could be implemented by a system rather than operators. Only after we had the first prototype running did it became clear we’d built a new kind of operating system, a meta operating system.

What’s unique is that a meta operating system has no APIs of it’s own, because the code runs in guest OS’s. Still, the meta OS controls resources, schedules processes and provides the user a console.

Realizing this was an important step and made completing the project easier. Realizing we had an OS meant we now had a template to refer to whenever we weren’t sure how to solve a problem.

Peter Nickolov, our President and CTO, was scheduled to speak at AjaxWorld on the 24th and 25th, but he’s been called out of town. So the audiences will have to put up with my blathering style of speech instead of his deliberate delivery.

The topics of Monday’s panel is still up in the air, so it should be interesting, but on Tuesday morning I’ll be talking about the development of our AJAX interface along with Stephen Quintero, one of our developers.

Rich Miller has a post about the Open Virtual machine Format (OVF) that a consortium of vendors submitted to the DMTF last week.

I’ve blogged about standards for utility computing a few times before, and I continue to believe it’s too early for a standard.

Standards are a double edged sword - a trade-off to gain interoperability in exchange for stifling innovation. Once sufficient experimentation in an area of technology has been conducted that agreement can be negotiated between competitors on market requirements, a standard can be drafted that delivers interoperability and allows innovation at the next layer. IMHO it’s way too early to make that trade-off. We haven’t adequately explored the possibilities in utility computing.

For instance, the OVF is intended to provide interchange of applications between services. To my knowledge, no vendor other than 3tera has ever demonstrated this ability. Our customers do it frequently. Therefore, we can document precisely what the requirements of an interchange format are and what services are required on both ends. OVF, unfortunately, is simply insufficient.

Even if 3tera published our interchange specification tomorrow and all parties signed on to it, that wouldn’t be the end game, because transfering an application between data centers isn’t the end requirement. It’s merely one important step. 3tera’s roadmap includes several major leaps in capabilities that will require significant extension of our current interchange format. And, as I mentioned, that’s where the negative aspect of standards comes into play. To adequately explain the need for those extensions in a standard body, 3tera would be required to site use cases where they could be needed - in essence to divulge our product roadmap to our competitors.

That said, if the authors of the OVF draft want input and are willing to embrace capabilities they can’t yet provide, capabilities that will set the stage for years of innovation, then we’ll be happy to participate.

I’m at Tier 1’s Hosting Transformation Summit for two days of meetings, presentationa and discussions with the leaders in the hosting industry. This is the second year I’ve attended and the Tier 1 folks do a great job of bringing together the senior execs from the major players.

Our CEO, Barry Lynn, did a great job on a pannel regarding utility computing. During the talk he made a rather controversial statement - that computing could eventually be made free.

The basis for his statement are calculations based on our user’s costs for operating online services versus traditional colocation. Projecting those calculations into the future there is indeed a crossover point - a time in the future when the cost of computing could be supported by some other monetization method such as advertising.

Sound ridiculus? So did getting applications for free at one point.

Marc Andreessen has an interesting post in which he attempts to catagorize internet platforms into three levels:

* A Level 1 platform’s apps run elsewhere, and calls into the platform via a web services API to draw on data and services — this is how Flickr does it.

* A Level 2 platform’s apps run elsewhere, but inject functionality into the platform via a plug-in API — this is how Facebook does it.

* A Level 3 platform’s apps run inside the platform itself — the platform provides the “runtime environment” within which the app’s code runs.

The attempt at catagorization is thought provoking, but has at least one oddity - Ning ends up in the same catagory as EC2/S3 and by extension AppLogic. IMHO this actually seems to call for a fourth level. I’m not sure whether I’d classify AppLogic as level 4 or level 2.5 based on Marc’s requirements, but you can read his post and decide for yourself.

Our founder, Vlad Miloushev, passed away suddenly August 30th. This weekend, his friends and family gathered to say goodbye and celebrate an incredible man.

Vlad was always a force to be reckoned with. Whether at work or in a social setting his energy was infectious. His ability to cut through hype and distill issues to their basic elements was amazing. Those of us fortunate enough to know him will carry his spirit with us always.

A memorial site has been created for folks to share their memories at www.vladmiloushev.com.