James Kuffner joins Toyota from Google Robotics – Business Insider

After the division’s tumultuous year, a cofounder of Google Robotics just left to work at Toyota.

Source: James Kuffner joins Toyota from Google Robotics – Business Insider

Sadly it looks like Google (or Alphabet now) has bitten off more than it can chew with its robotics division. Multiple leaders of the group have moved on, and so far no large public project has been shown. Please Google, PLEASE learn the lessons that Microsoft had to learn the hard way in the 90-00s.

 

Hosting Providers – You get what you pay for.

A tale of two hosting providers

I had to change away from my old hosting provider, cloudatcost. They were not keeping up with their promised service, and lots of user have complained about them. Despite 4 support tickets and 5 months of complaints, service is unchanged (more on that later). Due to this, I have switched over to RamNode (ramnode.com), and so far the difference is night and day (or perhaps 3G and gigabit).


Apples to Oranges

Always a good idea to say more than just “it’s better”, hard numbers help make decisions in a price/performance market. A great place to start is Serverbear.com. You can download and run a standardized set of Linux tests on any server, and automatically see the results online. Below we can see that although CloudatCost provides just slightly higher benchmark performance, it suffers by 2 orders of magnitude or more in all other categories. This test was done several times, and we can see that CloudatCost is consistently under-provisioning both the network hardware and the disk resources of their instances. On one of the support tickets I had with them, they claimed a “faulty router in the network”, but no improvements were ever seen.

SVZ - Hosting Providers

Cloudpro - Hosting Providers

Raw performance stats in a snapshot are ok to start, but how do you test reliability of service? pingdom.com. Though pingdom has recently cut back on the services they offer for free (gotta make a living), they still offer a very good way to track server outages and issues. Here we can compare historic data to see how often a service goes down, how fast it responds, etc:

cloudatcost uptime of 96%, outages every week. - Hosting Providers
cloudatcost uptime of 96%, outages a few times every week.

And here is a zoomed in view of the last few days:

Notice a Difference? - Hosting Providers
Notice a Difference?

Looking at the second graph, if you had to guess, when did I switch over providers? Fairly obvious, when the response time became smooth and fast, and the outages stopped.

One more quick note, a good online tool to hammer your website with simulated load to test for issues: loadimpact.com

Self Regulation (sorta…)

After more than 5 months of support tickets and public complaints, nothing changed. Cloud at cost never hit advertised up times, speeds or performance, though they do have a (mostly working…) online interface, so that is nice….

Here in Canada, we have the Better Business Bureau, a group dedicated to helping customers and business establish and report on breaches of trust. I submitted a formal complaint to them about the lack of service. This is what I got back after a little while:

Complaint Case: 1329069
Business Name: Cloud at Cost

BBB has made several attempts to contact the business regarding the above referenced complaint. We regret to inform you that we have not received a response from the company. The case has been closed as will be reported as UNANSWERED.

Despite all this, cloud at cost continually run “80% off sale” “Black Friday Sale” (all emailed to my account) and get new customer sign ups.

As much as I love to support local businesses, their service at this point is essentially a scam, as they deliver a fraction of the promised service.

DO NOT USE!

 

  • Disclaimer: I have not been paid by anyone, nor received any benefit from this info. This is just my own info gathering, done using my personal funds.