SaaS or SALSA?
I think the IT industry, more than most other industries, may be with the exception of Wall st., is FAD driven. Take a look at SaaS. According to analysts (I do not consider myself to be one) this is the best thing to happen since sliced bread. It has the following benefits
1. Reduces Time to Market
2. A snap to Deploy
3. No Infrastructure to Manage
4. No License to Manage
5. Pay-as-you-go subscription model
6. Why pay for all the custom features when you do not really use it e.g. SAP
Sounds great so far. So whats the problem here?
Lets start with why anyone uses software or IT to start with. I think, there are two reasons:
1. It is the cost of doing business, the price of entry if I may
2. It is used for gaining business advantage
Let's say SaaS is adopted as widely as predicted, i.e. every mom-and-pop has CRM and whatever else that is available, that would be like everyone in business using Word, Excel and Email. It just becomes the cost of doing business. You cannot use hand-written notes or slow mail anymore. Historically same things happened with the Typewriter, the Electronic typewriter, the Fax machine, Photocopier etc. SaaS is no different. So what?
No Doubt SaaS offers the above stated benefits, but, there are certain issues that could inhibit its adoption beyond a certain segment. So what is the inhibiting factor? It is not customization or performance or accessibility or scalability.
Let me put it another way. Think of SaaS as a being equivalent to a Power plant. Instead of all of us running our own generators, we get electricity from a centrally located power plant. We have created a Grid based distribution system to increase the reliability. In spite of all this we have blackouts. Then, I'm sorry to say this, power plants become terrorist targets. If we all had our own power plants i.e. little generators this would not be an issue, but we loose efficiency.
So what do we do? We use Generators or Flashlights as backups to the Power grid. Depending on the criticality of the application i.e. if it is a surgical facility in a hospital or if it is a data center running mission critical applications we have UPS systems. So what is the relation to SaaS and what does that have to do with SALSA?
SALSA is an acronym that I just created. Just like that! Software As a Local Service Appliance. There are two purposes or implementation possibilities for this concept. It is easier to explain and understand this using an example. So here it is:
1. Salesforce.com application as an Appliance i.e. Salesforce SALSA
2. Salesforce.com Appliance backup to Salesforce.com SaaS i.e. Salesforce SaaS with SALSA high availability or Salesforce.com CONDOM (Common Domain - suggesting that it is one application but with local availability protection)
Why would you do this beyond the above stated reasons i.e. UPS backup, Local Power Generator analogy.
As good as oats is for ones health, we need to package it, add sugar and honey and make it enticing for people to consume, hence we have Cheerios. The benefits of a concept/product alone is not enough. Packaging brings wider acceptance and adoption i.e. revenue and profitability. In the case of SaaS there are corporate cultural barriers to cross. Many Medium and Large Enterprises care about control and data security. If data is not in their own sight or under their clear oversight they are not going to want to implement it. Then there is the issue of backup, restore-on-demand, storage, regulatory requirements, compliance, audit etc. to deal with. Saas may provide all this, but can do it only as much as a Powergrid provides electricity. If we trust a decades old technology like Power Transmission to that extent (Blackouts sure do not make it any easier to trust), there is no way Large and Medium enterprises are going to trust SaaS. SALSA just makes this easier.
The data replication technology is mature enough to support the CONDOM implementation suggested above, if needed. This makes the process of continual update of the software easier. If it is just a plain SALSA implementation, upgrades need to be worked out. The good thing is, if someone does not want to upgrade, they do not have to, making the adoption easier. In the end what is important are the benefits of SaaS as mentioned at the beginning, not where the software physically resides. People buy a product for its benefits. We use public transportation for its benefits, but still like to own a car, unless you are poor or just do not have the need for it (Just like the current SaaS adopters and that is the point)!.
Why not?
Friday, March 9, 2007
Tuesday, March 6, 2007
Horizontal or Vertical? Its Neither. I'm just tired of it.
If there is one thing that history has taught us, it is, things are never just one way or the other. It is always a combination of this and that. Take cars for example, you can either have a true sports car like Lotus, Ariel Atom, Zonda, some Porsche's or a excellent luxury car like Bentley, Rolls-Royce or even Maybach. Your choices are pretty limited at the extremes.
Then there is the vast middle from Acura to VW. This is where the money is made. So what does all this mean to being Horizontal or Vertical?
In the world of computer hardware, there are two camps - the infinitely horizontally scalable camp and the bigger is better camp. Many online companies like MSN, Yahoo, Google etc. built their computing capacity using what are known as commodity servers - 1000's of them. Meanwhile the large businesses have few very big boxes, traditionally called the Mainframe.
These two approaches seem like extremes. i.e. similar to the cars example. There is not much money to be made at either ends, in comparison to what can be made in the middle. In addition, the pendulum swings from one to the other depending on the 5-year period you pick to argue in the last 20 years.
Horizontally scalable systems started because of a need - a need to create a large computing environment for the least amount of money. Over a period of time, one is going to realize that this is like transporting goods with a bunch of minivans hitched together and from a management and flexibility perspective its a nightmare to navigate. Especially when once faces hardware obsolescence.
Vertically scalable systems on the other-hand started by someone - usually a large corp - wanting to gain a business advantage over competition. Money was no object then. It is like a Freight Train with a lot of lugging capacity but can go only on rails and not as flexible.
But what most companies need are pick-up trucks. So there is not much to gain financially in a trail of minivans or Trains on tracks. But we sure need them both for specific purposes. Moreover, over a period of time, the Minivan trail is going to realize that it is much cheaper and lot efficient to run trains for certain purposes.
So high-end hardware manufacturers like HP, IBM and SUN should not change their strategies a whole lot like for e.g. what SUN is currently doing. That would just cannibalize their own future at the expense of some short-term gain. This makes them lose focus. Get them into unfamiliar territory and make them struggle. So please define your competence and try to make the best of it. That is how you outshine and outlast others.
Companies like Verari and Rackable come and go. They may even make a quick buck in the process. Their business models are not sustainable in the long run. Currently the pendulum is swinging towards the horizontally scalable end of the spectrum. It sure will swing back - give it 3 to 5 years. Quick money is made by keeping the markets volatile, not focusing on the long term. I'll be around and hopefully make an update to this article - one way or the other. Why not?
Then there is the vast middle from Acura to VW. This is where the money is made. So what does all this mean to being Horizontal or Vertical?
In the world of computer hardware, there are two camps - the infinitely horizontally scalable camp and the bigger is better camp. Many online companies like MSN, Yahoo, Google etc. built their computing capacity using what are known as commodity servers - 1000's of them. Meanwhile the large businesses have few very big boxes, traditionally called the Mainframe.
These two approaches seem like extremes. i.e. similar to the cars example. There is not much money to be made at either ends, in comparison to what can be made in the middle. In addition, the pendulum swings from one to the other depending on the 5-year period you pick to argue in the last 20 years.
Horizontally scalable systems started because of a need - a need to create a large computing environment for the least amount of money. Over a period of time, one is going to realize that this is like transporting goods with a bunch of minivans hitched together and from a management and flexibility perspective its a nightmare to navigate. Especially when once faces hardware obsolescence.
Vertically scalable systems on the other-hand started by someone - usually a large corp - wanting to gain a business advantage over competition. Money was no object then. It is like a Freight Train with a lot of lugging capacity but can go only on rails and not as flexible.
But what most companies need are pick-up trucks. So there is not much to gain financially in a trail of minivans or Trains on tracks. But we sure need them both for specific purposes. Moreover, over a period of time, the Minivan trail is going to realize that it is much cheaper and lot efficient to run trains for certain purposes.
So high-end hardware manufacturers like HP, IBM and SUN should not change their strategies a whole lot like for e.g. what SUN is currently doing. That would just cannibalize their own future at the expense of some short-term gain. This makes them lose focus. Get them into unfamiliar territory and make them struggle. So please define your competence and try to make the best of it. That is how you outshine and outlast others.
Companies like Verari and Rackable come and go. They may even make a quick buck in the process. Their business models are not sustainable in the long run. Currently the pendulum is swinging towards the horizontally scalable end of the spectrum. It sure will swing back - give it 3 to 5 years. Quick money is made by keeping the markets volatile, not focusing on the long term. I'll be around and hopefully make an update to this article - one way or the other. Why not?
Friday, March 2, 2007
Why Google invests Billions in building Data Centers?
Where would Google go next? and Why?
Google's CEO comes from a company where he worked on the vision - The Network is the Computer. Now he has the money to prove it. So what does this mean? Let me start by talking about Google's investments in the past few years.
Google has been investing in large data centers around the world. It has been investing $1B a year by conservative estimates on building these data centers. As far as it is publicly known it is not buying hardware from any of the major vendors -i.e. SUN, HP, IBM, DELL. There are reports that Google is making its own motherboards and servers through contract manufacturers in China/Taiwan. With that type of a cost structure, $1B would buy a lot of computing power that can only be compared to what NASA and Pentagon has. Why would anyone spend all this money in building massive data centers that is geographically diverse? Its not all certainly for making its search better or to get more advertisement dollars. It does those just fine with what it already has.
Now let us look at what Google has in "Beta" --> http://www.google.com/intl/en/options/. What is the underlying theme of all that is available? Put together What are they capable of doing? Case in point - Information access on demand - Search and gain access to Video (Google Video, YouTube), Library of information, Research papers, patents, maps, shop (froogle), mail, productivity software (spreadsheet, document editor, calendar), Comment on (blogger), Chat with (talk), Organize (Groups), Interpret (Translate). And have access to all of this from anywhere in the world as wireless technology becomes more accessible, which it is. What more could anyone want on a day-to-day basis to keep themselves busy/entertained/be productive? Do all this faster? That's why there is Google Web Accelerator! But there is something missing. What is the connection between these seemingly diverse applications, build-up of computing capacity, Google's vision of "do no evil" and Schmidt's "The Network is the computer"?
Well if you can do everything that a computer does from anywhere in the world through a wireless network connection and a very simple and affordable Google "device" - would that mesh well with everything in the previous paragraph? It is easy to phantom that the productivity of the entire human race would increase over a period of time? I think so. Not just that, Google can leverage all of this and sell us more through targeted advertising which they are already very good at!
So, to truly realize this "The Network is the Computer" vision along with "Do no Evil" (by providing affordable access to productivity tools to the masses) and expand on what they are already good at (search, advertising) you need computing power. If a billion people need to run spreadsheet, document editor, video streaming etc, you need to be near them to serve the request made to that application over a wireless network. For this you need to invest many billions in geographically diverse and redundant data centers.
Get ready to truly elevate the productivity of human race. What Microsoft did to a niche, Google, I think, wants to do to the human race.
Google's CEO comes from a company where he worked on the vision - The Network is the Computer. Now he has the money to prove it. So what does this mean? Let me start by talking about Google's investments in the past few years.
Google has been investing in large data centers around the world. It has been investing $1B a year by conservative estimates on building these data centers. As far as it is publicly known it is not buying hardware from any of the major vendors -i.e. SUN, HP, IBM, DELL. There are reports that Google is making its own motherboards and servers through contract manufacturers in China/Taiwan. With that type of a cost structure, $1B would buy a lot of computing power that can only be compared to what NASA and Pentagon has. Why would anyone spend all this money in building massive data centers that is geographically diverse? Its not all certainly for making its search better or to get more advertisement dollars. It does those just fine with what it already has.
Now let us look at what Google has in "Beta" --> http://www.google.com/intl/en/options/. What is the underlying theme of all that is available? Put together What are they capable of doing? Case in point - Information access on demand - Search and gain access to Video (Google Video, YouTube), Library of information, Research papers, patents, maps, shop (froogle), mail, productivity software (spreadsheet, document editor, calendar), Comment on (blogger), Chat with (talk), Organize (Groups), Interpret (Translate). And have access to all of this from anywhere in the world as wireless technology becomes more accessible, which it is. What more could anyone want on a day-to-day basis to keep themselves busy/entertained/be productive? Do all this faster? That's why there is Google Web Accelerator! But there is something missing. What is the connection between these seemingly diverse applications, build-up of computing capacity, Google's vision of "do no evil" and Schmidt's "The Network is the computer"?
Well if you can do everything that a computer does from anywhere in the world through a wireless network connection and a very simple and affordable Google "device" - would that mesh well with everything in the previous paragraph? It is easy to phantom that the productivity of the entire human race would increase over a period of time? I think so. Not just that, Google can leverage all of this and sell us more through targeted advertising which they are already very good at!
So, to truly realize this "The Network is the Computer" vision along with "Do no Evil" (by providing affordable access to productivity tools to the masses) and expand on what they are already good at (search, advertising) you need computing power. If a billion people need to run spreadsheet, document editor, video streaming etc, you need to be near them to serve the request made to that application over a wireless network. For this you need to invest many billions in geographically diverse and redundant data centers.
Get ready to truly elevate the productivity of human race. What Microsoft did to a niche, Google, I think, wants to do to the human race.
Subscribe to:
Posts (Atom)