Search This Blog

Wednesday, February 22, 2006

‘Fast Convergence’ – Oxymoron ? The case of Gmail-Talk

On one hand there is a dream of complete convergence. On the other hand there is a hope of simple and fast. I wonder if both can co-exist (especially the fast part). A few weeks ago, Google announced that their ‘Google Chat’ feature is now integrated with Gmail.

Hooray ! Really ? I wondered if one really wanted a browser integrated email system that could also chat. After all, chat is real time, email is semi-realtime at best. In addition, I got pretty tuned off seeing that Google automatically showed me online each time I checked email. Fortunately, it still provided the “standard without chat view” (see image below) at the bottom which I switched back to.

I then thought about speed – my main attraction towards google was simplicity & speed , not convergence. I use a dial-up very often and have often complained that is simply not dialup friendly. I also travel very often and use my phone as a bluetooth modem – forget broadband speeds there. I hate having to wait for the image slices to load up – can’t we just have a simple non fancy driving directions page ? (I use the old yahoo maps often for the same reason – not the new beta one, when travelling).

I also noticed that google has a new beta toolbar our – allows me to add many buttons, the toolbar search box automatically expands as I type, suggests words as I type (which I could fortunately turn off). But guess what – it was far more sluggish in operation (not search) than the old toolbar. So I rolled back to the old one. However, I guess this is only a matter of time till the new one becomes defacto.
But back to gmail-talk:
So today, I pulled up the very cool Fidder HTTP debugging tool and ran ‘gmail in standard mode’ and ‘gmail with chat mode’ to see the difference.

Here is the summary after clearing cache for each mode (more detailed shots later)

Click on the table for a larger version:

That is quite a magnitude of difference.

Here are the fiddler detail screen shots for those interested (click on the thumbnails for larger versions)Without Chat:

With Chat:

The same holds true for Google desktop 3.0 – its bigger, does more, integrates better and is sluggish compared to the leaner predecessors.

So again, is ‘Fast Convergence’ an Oxymoron ?


Effective Presentations

Sorry for the lack of posting the past few weeks - I was/am busy designing an IMS network for an agressive carrier (yeesh !)
Some tips I wrote up a long time ago for making effective presentations.
  • Always have an Agenda slide – and at the beginning ask if the customer would like to change the flow

  • Before you start, ask the customer for a time-check so you can best fit your presentation according to the time slot

  • Americans in general are very open conversationalists – an (appropriate) joke here and there or a side conversation to make a presentation more interesting is taken well

  • Avoid ‘speaking loud’ – many people tend to switch on their internal boom-boxes when presenting. It really sounds like those looney Americal Idol participants who think screaming while singing shows tone control.

  • Always keep your presentation interactive – if you see no one asking you questions, more likely than not, you are boring those who are listening

  • Try and keep your presentation short at the first level (5-7 pages for corporate, 10-15 for technical). Remember that customers always look at your page count to see how much is left, especially if it’s a boring presentation. It is very hard to retain interest after 20 pages. At the end of the shorter presentation, ask the customer if they want to deep dive and then take them through more details.

  • Always try to assess who is interested in which part of your presentation. Be flexible in modifying your flow. Being flexible in presentation flow is probably one of the most important aspects of an effective presentation. What works with one customer may backfire with another and you must switch gears to save the day

  • Know when to shut up – if you are presenting to a customer in a group, with more than one person presenting, do your job and then don’t interrupt the other person with your own perspective. This lack of judgement is often shown by senior people when more junior people in their organization are presenting – they butt in with ‘let me put what he said in perspective’ - a bit of perspective helps, but instead of acting ‘one up’ in front of the customer, rehearse the presentation before with your colleagues


Tuesday, February 7, 2006

Be a ‘Fon’-ero. Why technology is not just about technology ( Peer2PeerWifi Sharing)

I received an alert from Fierce Wireless last week that Google and Skype have recently invested in a startup company that offers a mechanism using which broadband ‘users’ can share their wifi connections with others.

The model is simple: You buy a ‘FON’ enabled wifi router ad plug in into your existing broadband connection. The FON router has special routing and authentication sofware which then allows this connection to be access by others.

This is not a new technology – wifi sharing is a very old thought (almost as old as wifi itself) and was never successfully implemented because of several issues:

a) Range - This only effectively works if the access points are ‘within connecting range’ of each other. If you have ever tried to surf the web on your neighbor’s hi-speed internet (now, now ! I am not saying I did !) you would realize that signal drops are very frequent and quite frustrating. On top of that, the Linksys boxes that FON uses (I guess Linksys is their hardware partner) don’t do a particularly good job with signal strength (In fact, in my experience, Linksys boxes are one of the worst in signal strength)

b) More than sharing - Sharing is not just about exposing your SSID. A true peer2peer Wifi sharing and routing protocol should take care of more challenging aspects such as failover of access points, IP change as a ‘FON’ user moves from one AP to other and other things

c) Security – Wifi also comes along with security concerns. Peer2Peer sharing technologies always have a concern of owners running detection software to read my passwords. For example, what if I connect to a network called ‘FON_naughtyboy’ which is nothing but an access point hosted by some teenage geeks who are sniffing anyone who joins their network ? Good end-end encryption is ofcourse the way to address this, but this being a P2P network, the basic assumption of ‘centralized trust’ is no longer there. We now need to trust that every access point in the FON network is well behaved.

d) Peer2Peer Applications vs. Infrastructure – there is a difference between sharing applications vs. infrastructure in P2P networks. When people download music using BitTorrent, or call using Skype, they are running P2P applications – not exchanging banking passwords or logging into personal emails. On the other hand, wifi infrastructure sharing means that shared infrastructure will be used for all applications, including personal ones.

So How is FoN addressing various P2P issues ?
a) Range and Reach – FoN has got Google and Skype as backers and Sequoia Captial and Index Ventures as investors. A strong team. Success of any peer2peer sharing mechaism completely depends on how many people sign up. FON makes sense for Skype: Skype has over 85 million users today and any effort that provides more un-regulated broadband access means more installs of Skype clients and more subscribers. FON also makes a lot of sense for Google. It is public knowledge that google is furiously working towards setting up high speed datacentres to offer broadband to the last mile to all users. In a recent post, I mentioned that they are already working with companies such as Tropos to offer mesh wifi. FON is another kind of mesh wifi where users participate in network buildout. In terms of range, FON has already started partnership with ISPs (such as Glocalnet in EU). here is a snapshot of FON access points in the US - the nice thing is as more users install more FON routers, this map will keep getting updated.

b) Security – what do you do when the underlying network cannot be trusted? You resort to end-end encryption with trusted hosts. Remember Google’s “secure access installer” ? Their logic is this – if you don’t trust the network, trust google. So, if people are concerned that p2p wifi mesh networks are fundamentally insecure, install this secure encrypted tunnel software on your PC – and the communication over the Wifi network will be ‘Google Certified’. Ofcourse, this is just a hypothesis – but this is how I see Google will expand its vision with FON.

c) Making Money - Fon is also working on releasing a billing network where 'FONeros' (FON users) can charge a fee if they wish when they share their bandwith. Several issues need to be ironed out here, but it is an intersting model.
d) Reliability - FON is playing a different paradigm here. It is about offering ubiquotous access points. You no longer have to 'search' for 'wifi enabled spots in starbucks' or airports. 100% reliability is not a requirement here. If an access point goes down, hunt for another. And it is not expected that a home accesspoint is being shared with thousands. Considering their limited range, it is likely that the number of users per access point will be much less than 100 !
e) Reducing Capital Investments - traditional centralized models like T-Mobile need a lot of upfront investments from T-mobile to roll out the wifi network. On the other hand, FON takes the captial investment away by chopping it up into very small and affordable investments for FON-eros ($25 to be exact). As more people participate, the network expands. And each small part of the network is 'self-managing' - if a FON user sees his device doesn't work anymore, he just walks to a bestbuy and buys another one for $25, at worst. And he plonks it back again into the network. No expensive centralized EMS solutions required - why ? because it is community driven.


FON is nothing new in technology. However, it has the right partners and the right investors. It’s partners already have a strong build-out network which FON can expect to ride. And when they use catchy lines like “Join The Movement!” and “Become a Fonero Today” – its already a marketing cult isn’t it ? Sigh.

Thursday, February 2, 2006

The power of blogging

I always read about how powerful the blogging community reach is, but never experienced it first hand, till now.

We began this blog in November 2005 (less than 3 months ago) as a ‘fun thing’ to do. Within 3 months, we have around 600 unique visitors reading our blog, with around 350+ returning visitors. Considering its short life-span, it is quite exciting to see the reach of this system.

This has basically taught us four things:

  • People actually read stuff if it caters to a certain vertical they are interested in (in our case, VoIP, Wireless, all things convergent)
  • Humor helps – our funnier ‘red tie yapping’ articles seem to be fairly popular
  • Sharing experience is appreciated – our red-tie posts on “to-code-or-not-to” and “being a good manager” are examples of some of the more widely read red-tie articles we posted.
  • Advertising revenue works – with google ads and similar
A Note about anonymity
We have received several emails on ‘who we are’ and ‘why we keep it a secret’. The answer is simple: Just for fun. We don’t really care if you know who we are. It is just a little bit more interesting using a nutty monicker. Infact, if you google around, it is very easy to atleast know who one of us is. No big deal. Like we said, it is just for fun.

Who are you ?
If you are reading our blog, how about telling us who you are ? How is this for a deal: You tell us who you are, what you do, and I will tell you who I am and what I do (if you haven’t found out already). Our email id is corporaterat AT gmail DOT com.
Want us to write about something ?
We are not claiming to be experts, but we do claim to be fairly comfortable in the areas of technology we write about, because we work hands-on in these domains and consult with clients who build solutions for this space. If you want us to write about a technology vertical within this convergence space, send us an email and let us know. If we do write about it, we would be glad to acknowledge you (if you want to).