LoudSpeakers

I was asked to put together a group of interesting folks to discuss issues related to, the oft abused moniker, Web 2.0 for the TiECon East event held in Waltham, MA on May 30.

I picked people that I thought were the most interesting and capable of having a conversation about the future of the Web and business models related to it. Thanks to the following folks bringing their insight to the panel:

Fred Wilson - A VC from NYC is a thought leader in this space with a wicked (we saw that word here in Boston) sense of humor to boot.
Don Dodge - The man with the coolest name in Tech is in BD with Microsoft’s Emerging Business Team. He always has insights given his vast operating experience.
Brian Balfour - Founder of Viximo, based in Cambridge, MA, is leading the charge with new ways to monetize social networks with digital goods.
Nabeel Hyatt - Nabeel runs Conduit Labs, a social gaming platform. You can find him sipping lattes at Open Coffee Boston and talking entrepreneurship.

And David Cancel, Lookery CTO and Founder, asked the interesting questions and steered the panel.

There’s insight into Twitter’s rumored recent $15M venture round. Apparently, it hasn’t closed and anything you read in the blogs was pure conjecture.

The conversation is great. But the video quality sorta sucks. Oh, and the sound needs to be fully jacked-up. But if you work in this space, this video should be of interest to you.

Here’s a quick list of things I want from Amazon’s EC2 Web Services offering:

1. Better I/O performance for machine instances because it’s currently less than desirable. Most applications are disk I/O bound not CPU bound. Being able to scale processing resources is great and it’s a wonderful thing to boast, but that doesn’t help make basic, practical services like MySQL any faster without a ton non-trivial application engineering. Restoring a large amount of MySQL data took 5 to 8 times longer on EC2 than it did on similarly equipped physical hardware.

2. I need to be able to call and speak to a human when things start acting up. I recently used the official Amazon 64bit Fedora Core image for some large scale processing that I needed. I restored a database with hundreds of millions of records and configured it appropriately. Apparently, the Linux kernel was leaking memory like a sieve (for no obvious reason) and Amazon told me to go pound sand when I posted to their message board. Kill the instance and start anew. In other words, scrap your work and start over again because we’re not going to help you. A good web host would have taken responsibility for their hardware and fixed the problem. Amazon let me down in a big way. P.S. Providing primary support via a message board is lame, lame, lame.

3. Graphical tools to manage instances. Don’t get me wrong — I like text console foreplay like the rest of the world, but give me web-based tools to manage AWS please. And don’t point me to some 3rd party service that wants to charge for this — it should come free with usage of Amazon Web Services. Seriously.

4. Dynamically allocate more storage to instances without some kludge FUSE based spaghetti-strapped hack. Seriously, it’s weak on too many levels to complain about. It’s almost embarrassing.

5. The ability to suspend an instance without killing it entirely. I need to put a configured instance to sleep without destroying it. There are times when I need to suspend an instance for some interval and then start it again —- I need a state between instance reboot and ec2kill.