Industry Observer

Natural Advantage created for the Service Industry by Cloud Computing

2011-08-04T21:04:00.000-07:00

The BIT Conference at UCLA Anderson had some great speakers and diverse mix from academia and industry. Two must see talks here
1. Terry Kramer on the wireless industry
2. Ravi Kalakota on Renovation as compared to Innovation

I was honored to be part of an industry panel. Here is my presentation

Natural Advantage Created for Service Providers by Cloud Computing

View more presentations from njuneja

Creating a Two Sided Market in Insurance using Technology

2010-06-28T11:11:00.000-07:00

Several people including me had big hopes on Pay-Per-Mile insurance. I even categorized this innovation as an Insurance 2.0 innovation and termed it as part of the next-gen insurance model. After all , Pay-Per Mile gives the ability for carriers to track your driving usage and you as the consumer are passed down the savings. Several agencies had research published claiming that the miles driven could reduce by 10% and a savings of $270 could be achieved. Carriers like Progressive and vendors like On Star launched their products with a lot of fanfare only to find that consumers did not really opt-in for the plan.

The problem, in my opinion, for the limited success of the product is that the benefits to consumer in pay-per-mile insurance product are not clearly articulated and more so, not proven. Knowing that there is a big brother watching you and that your costs could go either-ways (up or down depending on the usage) makes the consumers shy off going to a pay-per-use plan. Instead a better approach for carriers could be to provide a free value added service in return for the customer's allowing the carriers to track the usage patterns. Essentially creating a two sided market where one feeds the other

So here is an example of how this could be achieved. Refer to the diagram below.

In my sample next generation insurance company model below(refer diagram), I am creating a two-sided marketplace between the Commercial Auto (Tow Truck ) and the Personal Lines Auto (Consumers). Each side is feeding the other and also deriving value by joining this new entity. Consumers -Like the ability to have an efficient tow truck dispatch capability on their cell phones. The Tow Truck carriers would not mind sharing their location data when they are shown that they get new business referrals via this technology.

Spring Roo and AppEngine

2010-06-11T13:42:00.000-07:00

I don't really get what springsource is trying to do with the Spring Roo product. Seems like it is trying to be the Java equivalent of the RAILS/GRAILS world; Fast and quick way to develop applications.

But does the approach truly work? I don't think so.

So I decided to look under the hood and build an application. I tried the much publicized Google AppEngine/ Spring Roo Integration. While I could build the scaffolding of the application very quickly, I did not see much value in what Roo did for me. Here is what it does :
1. Builds the Object to Database connection layer with finders and getters based on just the Entity Object definitions.
2. Builds a generic GUI for Data Entry and View.

Now - if you are in the RDBMS world - there is probably some value to automatically creating the scaffolding around capturing the Data Access layer with a tool like Roo. But , if you are working with AppEngine , I dont see the value. Building the Data access layer in AppEngine is super easy already there is not much other tools can do here. You just define your objects with JDO/JPA annotations and you are done. Why would I make comprises in the choice of a programming model when I don't get much from Roo's efficiency. Note that Roo uses JPA only with No DAO layer and uses Aspects.(not everybody in the world likes Aspects including myself)

I liked the idea of building a generic gui for your object model. If this idea would work with any Object Model I create. i.e not something that was generated via SpringRoo then I would be interested in it. Since the gui is more of a testing tool for your Object Model - there is value in having such a gui as a backdoor entry to your data. But the GUI cannot be used even as a starting point for true real-world applications.

IMHO - the only way I see Spring Roo fit into my world if it targets its positioning as the "Microsoft Access for the Cloud" .Afterall , what MS access gives you is an ability to define Tables(Entities) , Edit Data Items and Report on them. Trying to target Spring Roo as a potential Web Application development paradigm does not fly with me.

The Business IT Alignment Problem -2009

2010-01-11T08:35:00.000-08:00

Here is an interesting piece of information I found on the Gartner website . Gartner published the 2009 results of Business and IT Priorities based on a survey of about 1500 CIOs.

I have taken the Gartner Table and mapped the business priorities to what the IT priority should be if the IT was truly aligned with the business needs.

Seems like the Business IT Alignment issue is not getting any better.

The Cloud is ready but the Enterprise is not

2009-12-08T14:15:00.000-08:00

Seems like the whole world is talking about Cloud Computing Adoption in the Enterprise. What surprises me more is that some folks with a good sized audience are saying that as the technology is ready and the cost cheap, Enterprises will start replacing their applications with Cloud based apps.

I don't think so.

Cloud Computing Adoption in the Fortune 50, the ones which have an IT budget > 1 billion , is going to be two step process. Step1) Outsource your IT Infrastructure to Infrastructure Managed services provider like CSC , HP/EDS or even HCL and Step2) Finally adopt Saas and IaaS on a case by case basis once you have some sort of a pay per use model working for the Enterprise with the Managed service provider.

You cannot just switch to SaaS applications overnight(or even in an year) , nor are most applications built for use in a virtual environment. I would not be surprised to find that most large enterprises have more than 80% of their applications that cant even be moved to VMWARE without modifications, forget about moving to cloud.

What I am seeing in the Enterprise is that they like the Pay per use model. This not only fixes their internal IT finance issues(like which BU pays what , what are the allocations to the department etc) , but also moves IT to an expense category as compared to Capital expenditure. As a result, I do see a strong push to outsource IT infrastructure , mostly to Managed Service providers with some cost savings , though not as much as what they can save by moving to cloud computing. Now the managed service providers,who typically have their own data centers, are pushing hard to at-least not have dedicated boxes for each application. i.e move to vmvare type solutions.

So my guess is that once the applications move to a virtual environment as a step 1 and the Enterprises start paying for the IT infrastructure in a monthly model , switching to cloud with solutions like CloudSwitch would not be very difficult. It would be like moving from one cable tv operator for your home to another. Although the switch is not 100% smooth like putting in a new appliance , you can still switch with about 2-3 hrs of work

If Oracle firewalls mysql

2009-12-04T08:34:00.000-08:00

Interesting post today here.

The discussions :

Ellison has offered to create a separate entity within a combined Oracle/Sun to house MySQL which would be “firewalled” off from the rest of the company, possibly with its own board of directors

I hope this does not happen. Not for the sake of Oracle(ORCL) but for the sake of mySql. I was earlier led to believe that Oracle will in-fact not kill mysql. Open source is not going away and unless they bought sun just to kill sun , the sun software suite is all open source , open source is strategic to oracle also. So , this made me believe that oracle will figure out a way of dealing with mySQL just as IBM figured out eclipse and rational studio relationship.

Now if they end up creating a firewall between mySQL and Oracle with a separate board of directors, then we should surely expect mySQL to die. Oracle will ultimately own mySQL and if it cannot monetize it without hurting Oracle RDBMS it is going to kill it(in time).

Update: dec 07 - Apparently Oracle says that the new york post report is false.

Update : Jan 17: More thoughts from Marc Fleury here http://www.thedelphicfuture.org/2010/01/save-mysql.html

With the recent trend to somehow treat OSS software as public property.(Refer Yahoo boss and the mysql stuff) it is only going to hurt the OSS movement. It will negatively impact the OSS acquisition appetite for Enterprises as they might have to bundle in the cost of Liability insurance when in acquisition talks.

Not a good approach.

The OSS route was developing into - I along with others write some software , get some adopotion and then redemption with an Acquisition . Now the redemtion part is questionable.

Enterprise IT - The future sales model

2009-07-13T11:58:00.000-07:00

In his blog post , Mike Speiser, argues that Small to Medium Business segment is the next growth engine for Software and that Enterprise IT Providers should look out for disruption in their space led by the consumer web revolution.

While I agree with the conclusion of Mike (refer my earlier post that made similar conclusions here) , I do not agree with some of his arguments.

Mike's key argument is that with cost of sales going down ,as a result of the realization of the vision of self service technologies and next-to-zero software distribution costs, it creates an opportunity for new startup's to take advantage of this and hence this is the engine you should be looking at.

I would argue that there is more nuance to selling into SMB than having great self service technology and a cost model based on SaaS. Although people think SMB and Consumer segment are similar as a result of the fragmented nature of the markets , there are actually some significant dissimilarities. SMB's make decisions in groups typically influenced by a large Enterprise facilitator. i.e Doctor's would buy office software that works with the files their hospital uses. Insurance agents will typically buy their Insurance carrier recommended IT software. In a lot of cases Enterprise IT vendors sell to associations and groups ,realizing the actual sale with the SMB business; as a result the assumption that you cannot have direct sales for SMB is not true. Assuming that having better self service capabilities makes their life simpler is also not necessary true. Their life gets simpler with more factors like better integration with their affiliates.

So to play in the SMB technology space you need to have a platform provider that aligns all the relevant players. Microsoft has been this platform provider untill now. People like SAP understood this and focused on the Enterprise Solutions that integrated the SMB into the Enterprise processes but did not enter the SMB space per se.

Google is trying to be the new platform provider as a result of the platform moving to the web. So the opportunity now exists for start-up's the bet on the success of this platform and hence be successful.

US governments view on Cloud Computing

2009-05-13T17:33:00.000-07:00

Here is an interesting viewpoint on cloud computing from an actual large scale buyer,the US govt, as compared to vendors who corrupt the definition of CC to align with their product/service positioning. The view points have varied from as diverse as - James Governor calling -"If there is a consultant in the room it is not a cloud" and Mckinsey saying that only the IaaS piece is cloud computing.

The US Govt issued an RFI for procurement of Infrastructure as a Service and articulated a very clear definition of the cloud and what they are looking for.

Here is what the Fed say's :
________________________________________________________________
"Cloud computing is a pay-per-use model for enabling available, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model promotes availability and is comprised of five key characteristics, three delivery models, and four deployment models."

They go on to describe the 5 key characteristics of
On-demand self-service , Ubiquitous network access , Location independent resource pooling , Rapid elasticity and Pay per use.

the Delivery models of SaaS , Paas and Iaas and the Deployment models of Private , Community, Public and Hybrid Clouds.
______________________________________________________________

I like their view point as it is an inclusive definition and does not restrict to Infrastructure only while also preserving the key tenets of CC like pay-per-use , location independence and rapid elasticity.

Seems like they are asking the right questions and are actually ahead of the Enterprise's in their thinking. The cool thing about this RFI is that it will potentially serve as a template for other Large Enterprises for procurement of Cloud Services. Funny how the govt has is becoming the leader in tech adoption in 2009 .

Summer of the Java Cloud

2009-03-26T14:51:00.000-07:00

For those widget company ideas (like here and here), which are typically worked upon by some developer as a hobby - paying $75 per month to Amazon,with the added overhead of managing your systemimages , seems too much.

I am to glad to see that java in the cloud with pay on true usage shaping up. After all Java is still the most popular language around. The options on the horizon include
a) Stax
b) App Engine for Java
c) Sun Cloud

On Pricing:
Most of the widget company ideas start off as a hobby and get a life if the idea has potential so to expect a developer to pay $900 per year(the base version of Amazon EC2) on something that is just an experiment is way to much. The starting price has to be free

It seems to me that google has the best handle on this market -The innovation sourcing market. They have a free entry point but then tie you down to their proprietry and great tools. Note that to a developer GREAT is a more important than propreitry and to the google business managers PROPRIETARY is more important than great. Not to say that one can substitute or replace the other. Ley us see what sun microsystems comes out with - They promise on interoperability but will they be able to give something great within the constraints of interoperability.

Note: I must also say that I recognize that sun's and google's customers are different and their Cloud strategy will have to align with their customer's needs. For Google it is all about recruting developers as franchise's of their advertising business. For sun it is really about Enterprise and Datacenter

Death of billion dollar opensource company.

2009-03-20T10:29:00.000-07:00

For those coming in late - IBM intends to buy sun for 6.5 Billion.

I read the news as the death of opensource as a revenue model, and validation of opensource as an important tool for running a software related business.

First of all let me start this blog piece on my definition of “what is an opensource company?” It is a company where you can associate the revenue of the company directly to the opensource product/service. While this statement sounds quite straightforward, it is in-fact not that clearly established viewpoint in the industry. People tend to bundle companies that exploit opensource technologies into the category of opensource companies. As per this metric companies like HP,Google IBM and even Microsoft all fit it as opensource companies because all are successful exploiters of open source technologies. I disagree. If someone like Capgemini was to say that they are in the People Business I can understand, because they make money from selling bill time of people. But if Google was to say that they are in the people business because people are there single differentiating factor - I would not agree. Google is in the advertisement business to any Financial Analyst.

To me companies like Sun Microsystems, Redhat , springsource and a ton of small companies are “opensource companies”.

Now going back to IBM-Sun. How are IBM and Sun really different from an open source strategy perspective?

IBM has a clear demarcation between for-profit-revenue generating products (like websphere , db2) and for-leverage-cost-only-opensource products(like eclipse , apache software etc), while SUN on the other hand has a phase-customer-in approach where all products are free to use and you pay for it only if you want an insurance policy against something going wrong.

This demarcation IMHO is the key to successful use of opensource. Opensource in itself is not a revenue generating business. It is a leverage and an execution strategy. It can be a very good market capture strategy (downloads leads to adoption), a very good component sourcing strategy (the way IBM uses Apace foundation), a cost effective developer relations approach and a very good distribution vehicle.

Today – Sun Microsystems and Redhat are the biggest opensource companies in the world. Sun software and services is more than a billion while redhat is still about 300Million. After Sun gets bought over by IBM there will be no billion dollar open source business and my guess is that someone like redhat might also get absorbed by Oracle /HP.

So my viewpoint is that future of an successful open source company is to get bought out to enable better monetization or somehow undergo metamorphosis in the way they think about revenue. To be able to monetize better you need to have relationship managers that talk to executives that actually make purchases. Yes downstream adoption of the technology in the organization is a good thing but it is not a differentiating factor for a purchase decision. You need to have a flavor of the product branded differently that has a mandatory cost(Rational studio and eclipse). Yes you might tolerate piracy but you need to make the customer obligated to pay.

Free for all approaches work for the first $50 Million revenue but do not scale.To grow to a billion dollar of revenue , you need to make customers obligated to pay after certain thresholds are reached.

How many computers does the world need?

2009-03-06T12:25:00.000-08:00

An interesting discovery for me today was - around 20 per cent of all the servers sold around the world each year are now being bought by a small handful of internet companies -including Microsoft, Google, Yahoo and Amazon.

And all this before cloud computing as really taken off. in-Aptly said by IBM CEO in 1943 and then aptly re-phrased by Sun Microsystems CTO in the present context - The world only needs 5 computers

The movement of 20% to 80% is where all the usage patterns and architectural patterns are going to change. As pointed out by Rashid in the FT blog - "every time there’s a transition to a new computer architecture, there’s a tendency simply to assume that existing applications will be carried over (ie, word processors in the cloud)."

A case in point to me is spreadsheets. The current-form-spreadsheet IMHO will not scale in the cloud architecture. With the massive amount of data available to business users in the cloud architecture - Datamining techniques like Clustering , Decision Trees , Network Analysis will be available as features and the current charting capabilities would really move to the world of visualization.

So I finally wrote a google gadget

2009-02-10T17:16:00.000-08:00

one of things I really miss in me from my collage days is the ability I had to solve math problems without a paper. I could solve problem in air because I remembered math tables upto 20 , generally remembered fractions till about 10/10 and also the pow() functions till about ^3.

So to try out the Google GWT API's what could have been better other than to write a Math Practice Gadget. Here you can click on the google button below to try it out

Thinking about the Google Apps Reseller Program

2009-01-14T14:43:00.000-08:00

Google seems to be serious now about making money from Google Apps . Read here

So the deal is that a reseller sells GApps at $50 per user/per year and gets $10 per user/per year as his commission. To the buyer the price is always $50 (irrespective of where he buys from Google directly or reseller).

So to make this a viable business proposition for a reseller - if you staff one full time employee to get this business going (fully loaded cost of about $100K yearly) you will need to sell 10,000 seats to break even. 10,000 seats is no small number. What kind of a business will need > 10,000 seats and which has such an hybrid- and semi controlled IT Environment that only SaaS based web apps with zero installation and server side control do a justice.

My answer: Organizations that have three layers in their business. i.e the consumer , the agent/broker and the Company. The middle layer in these types of businesses has been the most difficult to manage from an IT Standpoint because it is typically a small business and a different legal entity but is associated with the Company .

Just ask the your Real Estate Broker or your insurance agent how they get their Century21 or ReMax email. The more you will drill into the economics of just a simple email address the more you will realize the opportunity for Google Apps type SaaS solutions

Cloud Vendors and why Amazon should buy Sun Micro

2008-11-14T10:56:00.000-08:00

I added one more slide to the my Walk in the Clouds presentation .

What this slide try's to do is provide a heatmap of the the possible strategy of different vendors in the cloud computing space.

Let me start with Google and Microsoft- Clearly they have a 100% overlap in their desires and they are going for the end consumers of IT - Create a platform , Provide Software Apps and Make your devices work all in the cloud. There approach might be different(i.e Google trying to enter the market via Consumers --> Small Business --> Enterprise and Microsoft going the other way round - Move their Enterprise Customers first --> Then go small business ) but they are essentially targeting top part of the IT Supply Chain.

Sun : Like always wants to be everywhere- Dont know if they really can be though. They have a software and platform strategy while also having a Baremetal and HaaS strategy. I would really question their J2ME Strategy considering that Android is probably going to kill J2ME.

HP - To me is a still the traditional big iron vendor. Focussing on Hardware and leaving it to Google , Microsoft and Amazon to fill the PaaS , SaaS stacks. Their view of the world - Be an arms supplier to the Enterprise IT Armies of the world.

Salesforce.com : Started off in the SaaS space for the Enterprise and will continue on that route. There is a strong case for Google to buy Salesforce.com when it really gets on the Enterprise IT bandwagon.

IBM : a pure play Enterprise play. They have three businesses - Big Iron(like HP) , Enterprise Software (Websphere , Db2) and Services. I think they will really focus towards the "Private cloud" model because it fits in quite well with where they want to be. An interesting thing about "Enterprise Software" Vendors like IBM and Oracle is that they not have a lock-in (from a technical standpoint) .i.e they generally adhere to standards and interoperability(which at this time is really a must to be playing in the Enterprise Software space) - their Lock-in is really in their notion of "On Stop Shop" - "Sales Process" - "Relationship Management". Dont be surprised if you start seeing a move from IBM talking about interoperability of Clouds.

Now Finally - Amazon. This has been the greatest success in this space. A wildcard entry they really realized the notion of cloud computing. But now that the market is established - what do they do next? My guess is that they either need to go up the ladder(ie compete with Google and MSFT) or go down the ladder(Compete with Sun , IBM and HP).Sticking in the middle limits their opportunity for innovation. The next set of Innovations in cloud will now either come from the orchestration of consumer facing technologies (Cell phone , PC , applications) as a continuum or it will come from basic innovation in hardware. As the economics of hardware moves from being distributed(commodity PC's) to centralized (server based ) who says that x86 and multicore will continue on its run. Also as the business model for going for the low end of the market with Ec2 is proven - It is only logical to assume that the next real innovation in data center technology will come from the big-iron vendors.

So at a Market Cap of 3 Billion Sun is dirt cheap and hence a good opprtunity for Amazon to downstream in the supplychain of Cloud computing

Insurance 2.0 : Googlelization of the Insurance Industry

2008-11-03T09:40:00.000-08:00

When I got to know that there was a suggestion at the Gartner symposium that google is the next big entrant insurance carriers should be scared off, the first though that came to my mind was - Balooney - why would google enter a totally unrelated business which has no leverage from its existing business - at best I though they could come out with insurance products for the cloud computing space - i.e data security insurance , open source indemnification(which is really an insurance business). It was not until later I realized that there were in fact a lot of synergies between Google and Insurance Companies.

What is Google's greatest asset? I would argue that it is - They know their customer too well. By knowing there customer I do not mean the traditional meaning implied by having a CRM , some excellent relationship managers etc. What I really mean is having the ability the drill down to each individual customer and knowing what is going on in customers life- what is on his mind(google search) , what kind of news is he/she reading(google news) , what does the customer watch(youtube), where does he generally travel and at what speeds does he travel (google Android) , What kind of friends he has (open social) etc.

If you ask an Insurance executive what is the key product differentiator for an insurance company? You will probably hear - Underwriting and Rating. And what makes better underwriting and rating - Knowing your customer very well. DMV records , Credit history , Claim history ,Demographic information and a whole bunch of other factors- most of it available as Pay for use data by data vendors like DNB, Choicepoint etc.

So needless to say that there is a clear leverage google has from its existing business to the insurance space. and getting into this space will further strengthen its dataset on each customer as it will get a different aspect of data on the customer.

What has been missing in the underwriting equation in today's insurance companies is knowing about an individual more than the bucket (category) he belongs to. i.e Per Customer underwriting and more predictive modeling based capability as compared to historical statistical analysis of categories. i.e would you be more willing to underwrite a type-a profile (no-accidents , no-claims in the past) but who has recently been searching on - "How to get a gun" and has been generally out of house between 11.00 PM and 4.00 AM (known via Google android) to someone with maybe one accident in the past one year. In todays model of underwriting the second example will be deemed as riskier which is probably not right.

Irrespective of whether google enters insurance or not I do feel that the next generation insurance company (Insurance 2.0) will have to be a Web2.0 company - that harnesses every interaction of an individual in the digital world (Searches , Movement - via android like technology , Relationships-via social networking ) to better its underwriting and rating capability and choose a better risk.

Google = Microsoft and Amazon = IBM

2008-10-30T10:46:00.000-07:00

Wow - how did I come up with the equation.

Here goes the story.
Google's strategy is what Microsoft's strategy was in the PC era.
- Start with bottom of the market (Consumers --> Small Business and then Enterprises)
- Create Lock in in some form or fashion. Google creates lock in with non-standard based technologies
a) Google App Engine(a totally google technology that no one else supports or has).
b) Android - intelligent way of creating proprietary technology while also leveraging the user base of Java without using J2ME.
On Android: The key customer for Android IMHO is small business. When all small business's try to sell their stuff via google(adsense , Location based advertising) - that is when google will make money. Android is really a play of getting more and more small business's rely on google for sales.

c) Google Apps - Hosted on the cloud with Google API's - once a small business gets hooked into this - the switching costs to something like office live is not going to make it worthwhile to move.

Google is clearly going after the small business now. Their aim - make it as simple as possible for getting small business onboard and don't bother of open standards , interoperability etc which generally come into an Enterprise Sale discussion.

Amazon IMHO is now really targeting the Enterprise. Although Amazon has also come to the market from bottom up(Startup's were their first customers) but now it is clearly going for the Enterprise

It is not a tightly integrated stack. i.e you can pick an choose EC2 , S3 , SDB or SQS independently and hence no lock in. There are some non-standard technologies like SDB and SQS but they are really at the edges of the offering.

Amazon in reality has an open offering. Use EC2 and if you do not like it you can quite easily switch elsewhere.

The key to Amazon's success is going to be Cost of operations which will be achieved by getting scale and being efficient while key to Google's success is going to be Customer Lock which will be achieved by releasing compelling technology to build a developer ecosystem.

So what happens to the Hp's IBM's and Sun's of the world. Lets See... I dont have an answer yet.

Let the real test of open source business models begin

2008-10-09T13:24:00.000-07:00

After reading about the dooms day in venture cap , I could not help thinking about the impact this might have to the open source world.

With things like zero-based budgeting and re-looking at the base fundamentals of your financial models – it is going to be very hard to justify investments with no direct link to revenue – i.e open source is going to come under a lot of scrutiny.

Besides, when every company is financially stressed out, there will be a lot of pressure on improving the productivity levels in the corporation and as a result the supply of free developers/testers that the open source world gets at the cost of the corporate world will now be constrained.

People do open source for several reason but from a business model standpoint – You do it essentially for two reasons

1. As a Market Capture strategy – Achieve ubiquity, trade revenue for market share. The strategy here is that make pennies on a large number of transaction as compared to dollars on a small number of transactions. The issue in the current financial environment is that if you have a product that has not yet reached ubiquity you will have a negative cash flow because you probably do not have year maintenance contracts.

Sun Microsystems is right into this category of open source players and I think has a mix of successful and un-successful open source products. Sun seems to think that it will benefit from the downturn , but I would be hard pressed to believe that some of its products like openoffice will not be questioned for further investments.

2. As a sales influencer to some other product or a service. A case in point is google - Google does a whole bunch of stuff that is free hoping that you would ultimately end up clicking its advertisements if you are somewhere in the google world. With the pressure on Google Stock and lets say a decreased add spending – what happens to projects like GWT would be worth observing.

Comparing AppEngine , EC2 and Caroline

2008-10-08T15:44:00.000-07:00

I gave a talk to the Los Angeles Java User group on the topic of "Comparing Google App Engine , Amazon Webservices and Project Caroline".

A lot of what I presented was really distribution of existing content and discussions. The only unique perspective that I think I added was the slide below.

The way I look at any player in the Cloud Computing space is one of the four buckets

1. Bare Metal - People and Process driven , essentially traditional style with some Hypervisor - the only service opportunity here is the traditional EDS type outsourcing that utilizes skilled labor and pooled labor.

2. HaaS - Hardware as a Service. My Definition: Programmatic Interface for Hardware Provisioning . EC2 fits in perfectly here. Moves up from BareMetal as it minimizes on "People Services" and focuses on Pooled hardware capacity and real-time provisioning of hardware. Typically billed by Clock time - offers a lot of flexibility in terms of choice of language runtimes.

3. PaaS : Platform as a Service . The issue of scaling has been abstracted out by the Platform and you have flexible services that automatically provision computing. Also refered to as Fabric . App Engine is an excellent example here. Another interesting point here is that you billing will typically move from Clock time to CPU cycle time - because there is no longer some instance you need launch.

4. SaaS : software as a service. Gmail , salesforce.com , ebay etc. You basically do not care what is happening beneath the stuff. You are the consumer of the software and expect the application to run

Here is the full Presentation.

Cloud Computing

View SlideShare presentation or Upload your own. (tags: ec2 appengine)

The coolest feature of chrome

2008-09-04T16:25:00.000-07:00

So - What do I like about Chrome . "Create Application Shortcut" - This is probably the reason Google created Chrome. Isnt it all about the platform.

You can create shortcut , put it on your desktop and open it directly as an application.
Now there is nothing new about this.

The cool part is that for Javascript heavy applications like gmail , Google Reader - it zooms. and atleast for google apps It seems like I do not sign in again.

I also like the application like look for chrome when opened through the shortcut.

Considering that Chrome's starts a new process for every tab - this use-case seems like the most used case for me , as pure HTML browsing is not cool in chrome , firefox is probably better .

So what does Google open source ?

2008-09-02T15:14:00.000-07:00

In light of the announcement by google to develop a new Open Source browser Chrome , I could'nt help notice that all open source products from google (GWT , Android and now Chrome) are client facing and essentially fuel for making Cloud Computing a reality.

Very eloquently said by Nick - that "the weakest link in the Cloud is the browser" and taking it to the next level of serving applications only brings them closer to making Google as THE platform for SaaS development.

Note that google has not open sourced any of its server side technologies like BigTable, MapReduce etc. and I dont think they would open source them unless these become commodities (with Hadoop etc) and there is cost leverage by open sourcing them.

This also make me think - why is Sun Microsystem (NASDAQ: JAVA) open sourcing its core revenue streams(software products). If their vision is really to make money from hardware and they view software as an enabler to hardware sales - I can understand. But they are not doing a great job at hardware also.

SOA adoption down from 2006 - Are you surprised?

2008-08-27T22:36:00.000-07:00

Not me. Having read the story and a very valid perspective from Joe McKendrick . I think a point that got missed is that : wasn't this expected?

To me it was. The reason at a macro level is that time and again we have seen that "Information Technology Industry " is the definition of a cyclical industry. Just when you thought that Rational Unified Process was the silver bullet for software development , Agile methodologies came about and challenged RUP. Just when you thought that the only way to build high performance websites was by purchasing high-end Unix servers , Horizontal scalability avatar was realized via google and showed that you can put together a ton of small PC's and run a web-scale application like that of google's(note that this might change again with red-shift).

The problem in IT and business is that just about the time when you can solve a problem , it actually moves elsewhere. SOA came into existence when the mantra for doing business shifted from "Think Globally and Act Locally" to "Think Globally and execute globally" . So executing globally required lot of governance , control and to some - excessive redtape. It lead to the creation of a new position in the IT organization - VP shared services , responsible for managing all shared assets by the organization.

Now the Mantra for business seems to be "Think Globally , execute globally and Collaborate Globally" - By collaborate globally I am referring to Crowd sourcing and open source techniques. The collaborate aspect changes the IT organization structure yet again. Instead of having a business head for "shared services" whose aim in life was to own everything common used in the Enterprise , The Lines of businesses now have the option to use crowd sourcing techniques and leverage SaaS type solutions to achieve their goals. When they do that there obviously will not be any synergies between the styles of consumption of services but clearly there will be a lot of JBOWS (Just a bunch of services) . As a result in my opinion the SOA concept of coming up with a master API that will be used by everyone is failing.

Why the Cloud Computing paradigm will enter Enterprise faster than you think?

2008-07-30T22:33:00.000-07:00

The discussion of adoption of the cloud computing paradigm in the Enterprise generally shifts towards outsourcing of infrastructure and the associated security concerns for an Enterprise. While the outsourcing part is an important part of the story that will have significant effects on the low end of the market (Small to medium business), my take is that it will not have a big impact on the Fortune 200 Enterprise clients.

What will probably happen to the Enterprise Clients is that they will end up building their own grids for internal consumption. You will start seeing vendor offerings like Sun sell -Project Caroline, that will probably be the key differentiators for the big iron vendors. (although Project caroline will potentially be the Sun Grid offering to the Small and midsize business’s - I can very easily see a situation where Sun Sales force goes to a client like Morgan Stanley and makes a sales pitch for Solaris and its hardware based on its Project Caroline based capabilities)

So going back to my original point about why do I think a case like above will come to the Enterprise faster than you think.

Most of the Large Enterprises typically have an infrastructure group that manages its data centers. This is how the provisioning of new hardware typically works in an enterprise

Applications VP - I need to unarchive some 300 GB of data and then use it for some analytics that I need to perform at least once every month.

Infrastructure Guy - 1GB costs about X $ and 1 LPAR with 2 CPUS is about Y$ per year. You need to multiply this by 5 years to get the ROI calculation for your project.

Applications VP - Wow ! Why is the cost 1.7 times Frys.

Infrastructure Guy- Well it is all the overheads; The Company needs to pay guys like us who ensure that additional storage is installed correctly and that your group adheres to all the norms we have established

Applications VP- OK (whatever! Since I do not have any options!) , when can I get it.

Infrastructure Guy- it will take 2-4 weeks after the purchase order is approved and quote submitted.

This is how it will work when you have a Computing Cloud running internally in a large Enterprise.

Applications VP - I need to unarchive some 300 GB of data and then use it for some analytics that I need to perform at least once every month.

Infrastructure Guy - Here you go , call this API for Adding Storage and launching an instance. You will be charged by the hour

Applications VP- Cool , I am charged ½ of what you guys charged me earlier and I have the ability to turn off my meter when I do not need to computing power.

Infrastructure Guy - Yes , They have cut down on group and all my buddies who did not have scripting skills and just PowerPoint skills have been asked to go. I guess our overhead is now 1.1 X as compared to 1.7X .Besides if you consider the savings you get by switching your computing off when not needing it , we are probably cheaper than frys.

So to sum it up- the fine grained transactions between the apps group and the infrastructure group will become coarse grained via development of API's and when that happens the labour costs will go down like anything.

Cloud computing : Its not just about the infrastructure

2008-06-30T17:34:00.000-07:00

So what is the value of Cloud Computing to me(someone who is not an infrastructure guy) ?

The important shift in my opinion is the change that will take place in the programming model for Enterprise Apps as a result of this new paradigm. The SDLC is going through a major change from Develop - Test - Deploy - Release as time/people sliced functions to Develop-Test-Deploy-Beta-develop-test-deploy-Beta2 to functions within the scope of developer responsiblities.

Imagine having your dev/test/prod environments in the cloud all driven through eclipse/netbeans and the movement between the environment's driven through course gained services like AWS webservices API that interact with your hardware to provision the application.

Imagine a library of widgets/services that will make into the Enterprise just has open source made its way in.It is really the re-birth of the Component Based Software taken to the next level.

Imagine treating things like -webscale computing , massive parallel processing- hadoop , specialized analytic capabilities , etc as commodities that can be purchased at any point of time for an application.

Imagine having a large number of small software vendors that thrive on delivering small specialized services/widgets to whom infrastructure is an operating expense and not a capital expenditure.

The above to me is the big switch. Besides the aspects of driving efficiency in infrastructure as a result of scale , the could computing paradigm will enable the building of a large number of software services players that will ultimately compete with the Application development department in the Enterprise.

Note: the link to Eric Schmidt's Bear sterns conference as an interesting observation of -"there will be small number of big players and a large numbers of small players in the cloud computing space"

Realtime Analytics for the rest of us

2008-05-16T12:03:00.000-07:00

My two cents added to the discussions here and here.

John's points about the DW purchase decision being pushed to the SaaS vendor and being less relevant to the enterprise (analytic application mid-market customer) is the key to the big switch. A similar analogy would be that when an Enterprise gets its electricity from the electrical company (the GRID ). All that it cares about is the SLA – Can the supplier meet my XXX Megawatt per hour demands at peak loads of YY, 24/7 and after that the pricing is the decision point. What kind of generators the electrical company uses (Hydro , coal , windmills) is an important decision for the electrical company but certainly not relevant to the Enterprise using the electricity.

So this analogy leads to a bit of soul-searching for DW appliance vendors like Terradata , Greenplum etc. “Who is their customer?” are they taking up the role of GE (who manufactures turbines , windmills etc) to serve the SaaS vendors OR do they want to be offering solutions at a level higher to the end consumer that ultimately end up using their appliances. I think a little bit of both and mixture of a lot more partnerships is probably what is going to happen.

It is also interesting to note two other trends that will shape the BI world.

Fragmenting of DB market to specialized Database
Availability of webscale level specialized Databases like BigTable , Hadoop/Hbase at very low entry points.

These trends will lead to a development of a longtail-type market for real-time analytics in an SaaS model (example – Recommendation service based on collaborative filtering, using Predictive modeling results during the underwriting workflow for approving a quote).The reason this will happen is because

a) These kinds of applications are more focused and can be performed in silos’. The whole concept widgets moving at the next layer of functionality and reuse.

b) They are also better served by vendors whose livelyhood depends on bettering the algorithms that power the analytics engines.

c) With things like Bigtable , Hadoop/Hbase exposed to the world at a very low entry point , all it takes is one guy to improve an algorithm and expose to the world as a service.

Update: May 22 ,2008 : I ran into one more Cloud computing Data analytics solutions which is another Column oriented database check it out at Vertica

Cloud computing and Datawarehousing

2008-04-09T11:57:00.000-07:00

Following my last blog entry , my brain continued elaborating on the thought of Cloud Computing Adoption in the Enterprise.

Having worked in the Enterprise Space for So long , I am hard pressed to come to terms with the notion that Enterprises will be willing to completely outsource their Information Management and IT infrastructure and more so in a constrained environment like that of Google AppEngine.

Amazon AWS with its ala carte is still a better option to Enterprises as compared to Google App Engine. As you can pick an choose what you want. I think the key is that any Cloud computing vendor needs to "FIT IN" into the Enterprise's Architecture . This basically imply's that more entry points you have to the cloud infrastructure the more use-cases you will have for Enterprise Adoption. So , it seems like Amazon has a better strategy for Enterprise Adoption. Another use-case for Enterprise Adoption is via an SaaS vendor case in point - Vertica . Vertica is a user of cloud infrastructure from Amazon. But the way the cloud is coming into the Enterprise Architecture is via an SaaS vendor.

Another use-case that came to my mind is the impact to the EDW world. With things like BigTable and simpleDB exposed , why would an Enterprise invest in highend Databases like TerrrData - why not use a proven scalable platform like BigTable to run your analytics. In any case you need to do your EDW work in house on separate machines from your core systems - So using an on-demand infrastructure for such needs makes sense.