Cloud Computing and Russia

With everything moving to the cloud, the demands on the the cloud compute infrastructure is going to be enormous - especially the cooling requirements. The higher the compute power, more the cooling requirements. Enormous amounts of energy would have to be spent just on keeping the infrastructure temperatures below certain values. This energy spent on cooling might in fact have an adverse impact on the "green footprint" of the massive datacenter in the "cloud".

There is some very interesting talk that has begun to take place in the cloud computing community. How about moving data centers to arctic countries ( maybe northern Russia, or Greenland/Denmark, Iceland, etc)?The energy required for the cooling would be potentially completely eliminated , thanks to the very cold climate. That is great for:

1. The cloud service provider
2. The earth in general ( "Green footprint")
3. The economies of the cold countries ( Surely , they would possibly love to invite the Cloud service providers)

This raises an interesting new economic paradigm akin to "outsourcing". There is a huge market potential that can be tapped by these cold countries. If the costs of moving and running data centers here is lesser than the OPEX of running the cloud service elsewhere ( which is really the key point ) , this would be a win for the cloud service providers as well. So a potential win-win situation.

This possible paradigm would possibly work out super for Russia, who is looking to improve economically , and who have an enormous "real estate" in the arctic. There is always some technology outsourcing going on here as well, even otherwise.

40GbE

Congratulations to Mellanox for announcing one of the industry's first 40GbE solutions. Its definitely the next big step in the evolution of Ethernet - and it has indeed come a long way from the conceptualization by Metcalfe.

Getting networking software working at 40GbE will be an enormously interesting challenge(Though the 40GbE adapter has the capabilities for running storage protocols such as FCoE obviously in hardware). The existing software TCP/IP stacks will definitely run into *severe* issues trying to get the best out of the huge pipe lying in front of it. TCP/IP was never designed with such speeds in mind. This is not to say that there are no solutions to the problem. The CPU speeds are increasing fast, but CPU speed increases are more of a linear trend - unlike the "leaps" in network bandwidth evolution.

The market adoption however might be an issue - considering that even 10GbE doesn't have the necessary traction it was presumed to have.

Anyways - good luck folks.

SaaS model for Outsourcing companies

With the advent of cloud computing and Software as a Service, what would be the implications on software outsourcing , especially Indian? The Indian outsourcing companies have reaped the benefits for over a decade , in companies world wide opting to lower software development and maintanance costs by outsourcing the development of many critical enterprise applications. 

Cloud computing is now offering the same benefits (potentially) to companies that had trouble doing in house software development. Does it mean that the outsourcing business model is bound for the decline in the next decade and will have Indian IT outsourcing shops scrambling to survive?

Not really. Consider the case of Wipro, which now entered the SaaS market with its own cloud offering ( partnering with Oracle). The firm has shown a willingness to adapt and assimilite new business models in the context of changing world wide technological landscape. This adaptation is more than just a good-to-have , and will probably decide the success and failure of many IT outsourcing companies. 

There are however some infrastrctural downsides for Indian IT companies entering the cloud computing space. The model assumes an uninterrupted availability of power especially to run massive hosted data centers out of a developing country such as India. Considering all reliability metrics needed to run enterprise apps, is this really feasible(yes, you can always have backup power for small operations, but for running massive datacenters, what would be the costs)? 

There are some questions here and I am pretty sure that the IT outsourcing companies are thinking through the issues. 

Passion and Practicality

"Entreprenuers tend to believe - "I have got my idea. I'll go with it until I die." But I advise them to take seriously the questions about whether their plan is irredeemably flawed, and whether they need to change what they are doing. Be diligent about failing fast so that you dont spend five years doing something thats just going to fail"

- Reid Hoffman ( With Alyssa Abkowitz, Fortune)

Pretty nicely put. Its very easy for entrepreneurs to go overboard with their convictions. Reality checks are needed from time to time and the dividing lines between passion and practicality have to be clearly marked.

Datadomain

Big consolidation news yet again from the datacenter space. Netapp buys disk based backup company Datadomain ( a competitor) for $1.5B, with a 40% premium on the closing price for May 20th. This is a clear case of further partnering of NAS and deduplication technologies. Netapp/Datadomain can now go against EMC.

Wolframalpha

I was curious for this release as something that would change the web as we know it - atleast sometime in the near future. "Computational Knowledge Engine" - wow - right out of an Asimov Novel it sounds like.

However, I was a little disappointed when it didn't even give me simple answers.

I tried:

"Which is the largest known Galaxy"

The Engine says:

Wolfram|Alpha isn't sure what to do with your input.

I then try:

"Which is the largest known galaxy in the known universe"

The Engine again says:

Wolfram|Alpha isn't sure what to do with your input

Finally I just say:

"Which is the largest galaxy"

Again the same output.

I know its still an early stage alpha - but I would have deifnitely liked something a little more sophisticated than the answer to , say, "What is the boiling point of water" for which you get an answer.


Btw, it didnt even give me the answer to :

"Who is the Chief Minister of Karnataka?" - Karnataka being my home state in India. It was however prompt enough to give an answer to "Who is the Prime Minister of India"

Long way to go me thinks - but the Idea is good.

TiECon 2009

I was there at TiECon 2009 , in Santa Clara , CA.


Here are some of the trends in technology I am seeing, at the close of the conference:

1. Clouds: Yup, they are everywhere. Pioneered( in some ways) by saleforce.com, there are now numerous vendors in the cloud bandwagon. The majority of the players I see are in the SaaS (software as a service) space. The SaaS players have now started offering PaaS(Platform as a Service) - exemplified by force.com, an offshoot of Salesforce.com. There are also players focussing on just the IaaS(Infrastructure as a service space) such as, Nirvanix(cloud storage) and certain components of Amazon EC2. There were companies talking about how they are seeing revenue improvements by building out apps on top of the cloud. "Cloud" is the new "Virtualization". The VCs too like the buzzword though there are claims by the tech press that the terms "cloud" is overused.

2. The Funding Scene: VCs are seeing a significant drop in follow on rounds. One of the VCs just said "Dont even think about it". The big delimma for companies now is what they would do should they be running out of a series A during this time. One of the VCs advised trying to work with existing Series A investors to keep the company afloat during these tough times. If you try to "sell out" there is no chance that the common share holders can get any returns at all - in fact , with high chance, the preferred will be under water too. One of the VCs mentioned that the big delimma for them is what to do with companies that have "significant" Series A/B ( tens of millions).

For early stagers, if you have "Cloud Apps" or "Mobile infrastructure/platform apps" in your story and have a reasonable customer base, maybe its time to pitch for a Series A. Be careful of the valuations though.

3. Building on top of new infrastructures: The infrastructure laid for the internet, gave rise to the bif search engine, the social networks , et al. The infrastructure laid in the late 90s for wireless gave rise to the mobile phenomenon and the rise of the iPhone, and all the wonderful apps. We now have an infrastructure of utility computing, where there is no need for setting up an inhouse hardware environment. We will probably see an era of great enterprise apps and business models using utility computing as the infrastructure.


Two great speakers in the show were Brad Smith ( CEO, Intuit) and Tony Hsieh(CEO, Zappos). One of the quotes I was particularly inspired by was Brad Smith's comment on the importance of business relationshops and parnerships in the new economy. He said, this was what his grandmother told him when he was a kid:

"We are all born angels with just one wing. The only way we can fly is by holding on to each other"

Great stuff!