(this is page 82 of 121)





Latest

These days everyone wants to grow a big cloud-based service with millions of users and billions of page views. Even Microsoft has got cloud-envy and wants to plunk down $50B to prove it. Unfortunately, most of us mere mortals have tons of trouble really getting these services to scale once we do get the users in the door. Things start to crawl, drives start to fill up, and before you know it you're Twitter with everyone biting the hand of the free service you've given them. Including this piece that argues for fronting Twitter with a proxy based of some sort of distributed hashtable thing.

While the caching/proxying thing has worked great wonders for most of todays super-scaled cloud services (YouTube, Facebook), I doubt it will continue to be the panacea for write-heavy communications applications like Twitter. Even the infallible Google seems to occasionally hit speed bumps with its venerable Gmail (which is not RDBMS-backed but is no doubt cooked from dilithium crystals).

It strikes me that the right solution to this problem is to revisit the notion of making distributed Internet applications work— but not the kind that have the service provider scaling across cages and datacenters. Instead what we need is to find models for using the "dark matter" of the Internet, namely all client computers connected to broadband connections and sitting idle for most of the day. NAT traversal issues and the mass-migration to laptops notwithstanding, these machines could be what is needed to help truly scale cloud-based services.

And even if all of those home/office PCs are too sporadically available, we could begin to rely on the growing number of personal accounts being created on VPS/shared hosting/Dot-Mac/whatever environments where each user is given some fractional part of a CPU and disk storage to do with as he pleases. This is why Wordpress's attempt at distributed Twitter, launched just a couple of weeks ago, is so interesting; most shared hosting environments provide a Wordpress install which means that with a little work (and some coordination between Twitter and the Prologue team), everyone can help take the load off.

There are "business model" reasons for why pulling off a hybrid implementation like this might prove difficult. After all at the end of the day, most consumer cloud services are valued by how many users (and more importantly user data and metadata) they are in possession of at the time of acquisition. Technical challenges aside, I am 100% certain that had someone pitched this approach to us at Tabblo back in 2005, we would have flat-out rejected it for all of the wrong business-model reasons. However, now that we are moving data centers though, I'm keenly aware of all of the terabytes of high-res image data that we must now shepherd across the continent when for most of what we needed, we could have done with a lot less (incidentally this problem is much worse for all of our bigger older cousins in the photo-hosting business).

Hopefully sometime soon we'll all get back to the original architecture of the Internet when it comes to this stuff— loosely coupled and distributed when it counts the most— and the scaling question will morph into one about writing large-scale distributed systems, a much more fun problem to work on.

Comments Permalink

Datastreams that matter

Fabulous piece by Tim O'Reilly in Radar on why Yahoo deserves to get swallowed by Microsoft. To borrow from Paul Graham, Tim seems to be arguing that Yahoo just never got with the "grain" of the web despite having bought awesome grain assets (Flickr, Delicious). The money quote here:

The other important characteristic of the winners, of course, is that they tap into a data stream that really matters. Owning network effects around consumer photos, for instance, is much less powerful than owning network effects around paid search. So one of the key questions we have to ask ourselves going forward is this: what are the major data subsystems of the future Internet Operating System. Location, identity (and social graph), search (and not just web search but also product search, in which Amazon has a very strong position) come to mind. In a lot of ways, finding the data associated with the old vectors who, what, when, where, and how is a good place to start.

While the jury is still out on whether the "social graph" belongs in this primal set of key types of data that you can build huge businesses around, I completely agree with the others that he lays out. It would be interesting to explore how some of the big Internet trends tend to interesect these different types of data: mobile computing, virtual worlds, custom manufacturing, and emergent online marketplaces just to name four.

Comments Permalink

Microhoo

First Google skids and then Microsoft comes in with an aggressive bid for Yahoo. Oh how we are coming to the end of this Web 2.0 party quickly.

I think it is a wonderful idea to put Microsoft and Yahoo together for a couple of reasons. First, despite what you may think about Microsoft, one thing they know how to do is build software at platform scale, something that Yahoo desperately needs to build the social platform they rightly seem to be intent on bringing into existence. Unlike Google and Facebook, the best elements to Yahoo's social platform were purchased and despite best efforts to bring them together under one Yahoo umbrella, they still feel like really different properties. Hopefully Microsoft can do the Office thing to Yahoo's best properties (Flickr, Mail, Finance, etc.) and make the whole thing feel well integrated. And in a bonus round they might also bring their whole developer-community effort to Yahoo's nascent attempts at building such a thing around Yahoo UI and a few web services.

The second reason why this is a good idea is that Microsoft can afford to take a much longer view on Yahoo's turnaround than Jerry & co. can under pressure from Wall Street. Just as they are nurturing the whole XBox platform, giving Yahoo room to breathe during the creation of its social platform could leave Microsoft with a great asset. Plus, the same could be said for a combined ad platform.

Finally, and despite the fact that this never happened with Hotmail, Microsoft-Yahoo could really deliver a seamless integration between rich client and web services. Every day that passes this becomes less and less relevant (as web browsers get better and people get used to the benefits of the cloud), but it's still going to be a while before we get back to ubiquitous broadband/4G, and in the meanwhile the combination of Yahoo's web savvy (including its recent acquisition of Zimbra) with Microsoft's emerging client platforms (.NET in the mainstream and Silverlight on the bleeding age) could conjure up some interesting brews. This was after all the idea behind the Adobe-Macromedia merger— which seems even more compelling in this case because it is not just two tools companies, but one of the greatest tools companies with a massive audience company.

Since I've been traveling I haven't had a chance to check on what the reaction to the announcement has been; however, I'm sure there are plenty of people who are using Scott McNealey's expression from the old HP-Compaq merger ("it's like two dumptrucks crashing into each other in the night"). While I can sympathize with the sentiment, we might pause to consider how the whole HP-Compaq thing has borne out— vaulting HP past Dell as the marketshare leader in PCs and notebooks and re-invigorating HP's entire PC business.

Might Yahoo not do the same for Microsoft's web businesses?

Comments Permalink

The day it went pop

I lived through one dot-com crash and all of the associated aftereffects, so I'm not psyched about the fact that the day of reckoning has finally come for this cute little kumbaya we've come to call Web 2.0. Sure on the surface it's just Google being slightly off their spectacular growth trajectory, but deep down we all knew this moment was coming. The wonderfully intoxicating "build critical mass and the rest will follow" buoyancy that started with Flickr and ended with Facebook's ridiculous $15b valuation is about to come to a close, and not so much due to one bad quarter of ad revenues, but because it just couldn't last forever.

Small companies struggling with product risk are now going to have to answer the revenue model question as well, and for good reason. Selling to Google— or even selling to someone else who is afraid that Google might buy you first— has just run out of juice. And unfortunately it has done so right at the beginning of a pretty scary set of perfect storm factors in the consumer economy: depressed consumer confidence, tons of credit risk, looming creep of inflation, and an inscrutable political outlook.

But what to do in the middle of all of this? Get somewhere where you can work on something meaningful with a 3-4 year runway. If you are at a startup, either get cashflow positive or raise a buttload of money soon. If you are at a big company, get on the longest lead time (but critical) project that you can find. Put your head down and just shut out the crap that is about to start flying.

It's all going to be good again— it always is. The challenge is that it may take until 2011 for us to get there. Until then, we can all stop worrying about getting rich and get back to grinding it out. Good times.

Comments Permalink

Learn to demo

David Beisel has done a great service for web ventures in the Boston area with his "Web Innovators Group" gatherings. Though this type of event is commonplace in the Valley, we are sorely missing it here and it's been a pleasure to watch it grow over the last year and a half from about 50 people to over 800.

One thing I continue to be struck by however, is the poor quality of the demos that are given at these events. It may be that David is explicitly selecting for non-venture backed companies (a little rougher around the edges), and it may just be the compressed format (though this being the week of the original demo conference, I'm not sure about the latter), but even when the companies on stage have cool and innovative products, the demos done WebInno generally leave a lot to be desired.

Since, giving a good demo should be secondhand to any startup employee (CEO, founder, hacker, product manager, etc.), here is my 3 step guide to coming up the learning curve quickly (though I don't claim to be an expert, I have given a lot of demos in all sorts of contexts and for all sorts of products in my career).

1. Write the story: before you start, open up your favorite text editor and write a one-pager on what you intend to show. Make it less laundry list and more murder mystery. If you don't know what progressive disclosure is, go read up on it, and use this technique to move your story line forward. The number one reason why I often want to put a stick in my eye during demos is because the presenter hasn't thought enough about telling a story and being entertaining. Make it relevant to the audience at every step— which often means skipping the techno-babble and contextualizing features for your audience.

2. Find your own Demo God(s) and study them incessantly: everyone loves Steve for this one, but there are many many others presenters worth watching (go look at all of the highly rated TED talks for inspiration). If you go to conferences, pay attention to presenters that can work with the crowd and watch how they keep cool and on message even when things are exploding. If you are lucky enough to know anyone whose demoing you respect, ask them for very critical feedback. I've been very lucky to have people like Adam Green to rely on for raw and uncut feedback. It hurts but makes you 10x more effective.

3. Practice, practice, practice: don't just think about what you're going to demo— go and sit in front of the mirror and do it. If you're going to be miked, learn how to use the AV system to your advantage. If you need the Internet for your demo, assume it will crawl and work around it. Practice it until you are so tired of hearing yourself that your ears start ringing. There is no such thing as too much practice. Trust me, even Steve does it.

One last thing: if you are not particularly good at this type of thing, try giving demos of products you haven't actually worked on. Textmate, OmniOutliner, Firefox, OneNote, whatever— just step through 1-3 for any one of those products assuming you'll have 5 minutes to convince someone that they absolutely need to start using whatever it is you are showing them. This trick helps to de-personalize your subject and lets you focus on your own performance.

There is a wise saying in raising venture funding: the demo seals the deal. Think about this next time it's your turn at WebInno.

Comments Permalink