Indian Software Industry

[hit&run]

@shade Has answered most of it. But the data center piece needs expansion -

Data centers while important aren't enough. To build a high scale app you need ready made services and infra running on top of hardware. Only a handful of providers have such wide capabilities - AWS (Amazon), Azure (Microsoft), GCP (Google) and probably Alibaba Cloud. Basically its the software and not (just) the hardware which makes them (big tech platforms) so useful. This is why Parler couldn't simply "get off Amazon"

Don't expect GoI or the sarkari employees at NIC to make something even remotely comparable. The closest in India is probably Flipkart's Cloud Platform - Flipkart doesn't use AWS and does most things inhouse (despite funding by Microsoft). Maybe they can convert it into a desi AWS. But there's no incentive for it. Yet

In terms of men hours and funds how long and what kind of investment we be looking at to create this software?

 

[shade]

This field is really hot right now, if Ambani tries he can pull another Jio in this field, but it also requires some serious top of the line tech and not only using that but also offering to customers. Which in turn would require huge R&D which Indian firms never want to do. IDK why?

The trader mindset here is only to go for the low hanging fruit/ez money, no appetite for high risk, high reward, and once they find some Kamdhenu Gau business niche, they will keep on milking it, atleast most of them, Mukes has expanded into Telecom and Retail from Oil refiing.

R&D needs huge budgets without any guaranteed results or rewards.

App stores aren't the real problem, the hosting sites such as AWS are the real issue.

No other hosting service comes close in terms of quality to price ratio. Without this no app can scale.

Yes, I'm currently paying 700 to takla for 2x Lightsail servers.
No issues from their end i've experience, and i'm now using since two years, paying since last year since first year is free.

 

[Cheran]

Thanks for very detailed response. I owe you few drinks after this. I will only have soft drink though.

My intention is to create an environment which is foolproof and can not be wiped out by big Inc.

Another requisite I may add is email hosting as well. When I joined GAB the email host I am using did not allowed confirmation email to to come through them. When I contacted GAB they told me that many email hosts have blocked us.

By Indinet I meant was a ring fenced internet within our boundaries which is controlled by us with kill switch in our hands. What comes in and goes out is controlled by us.
Social media app with only Indian citizens on it, running on our own Indinet, our own IE, our own websites for video, audio and picture hosting, own news portals (Doordarshan), fact checkers, own email hosting, maps, Indipedia, own app market place all of it conducted on our own data servers.

I think this is the whole eco system which we must establish otherwise, no offence to koi app but they seems quite vulnerable to me for the time being.

Practically difficult.

Almost all the known brands like Cisco (Routers/Switches/Wireless controllers/access points/NG firewalls/ISE/virtual solutions etc), Microsoft, FireEye, Avaya, Barracuda, F5, Fortinet, Sophos, Symantec, Brocade, etc ,, (you get the drift) are subject to wokeism & US NSA influence.

 

[shade]

In terms of men hours and funds how long and what kind of investment we be looking at to create this software?

Open Source Cloud Computing Infrastructure - OpenStack

OpenStack is an open source cloud computing infrastructure software project and is one of the three most active open source projects in the world.

www.openstack.org

There is OpenStack which can serve as a starting point, obviously nowhere in quality and performance as the Big 3.
Regardless the Sarkari organizations like NIC and CDAC cannot do this without management and "culture" changes.

Money and man hours I have no idea, but since like less than 10 companies are in this field, you can understand what big bucks salary taking people, hardware, and infrastructure is needed

 

[Nicky G]

Thanks for very detailed response. I owe you few drinks after this. I will only have soft drink though.

My intention is to create an environment which is foolproof and can not be wiped out by big Inc.

Another requisite I may add is email hosting as well. When I joined GAB the email host I am using did not allowed confirmation email to to come through them. When I contacted GAB they told me that many email hosts have blocked us.

By Indinet I meant was a ring fenced internet within our boundaries which is controlled by us with kill switch in our hands. What comes in and goes out is controlled by us.
Social media app with only Indian citizens on it, running on our own Indinet, our own IE, our own websites for video, audio and picture hosting, own news portals (Doordarshan), fact checkers, own email hosting, maps, Indipedia, own app market place all of it conducted on our own data servers.

I think this is the whole eco system which we must establish otherwise, no offence to koi app but they seems quite vulnerable to me for the time being.

Look at the entire stack. It's goes even deeper.

Even this won't be completely invulnerable, sites can be blocked at OS level, even chip level if you are really serious. So go build your OS to compete with Android/iOS and chips that cell phone manufacturers could use Intel/ARM etc.

The hardware for a internet is completely separate.

This ecosystem was built over atleast 2 decades, you'd need war time like funding and focus to achieve it in 10 years.

I don't why but I have recently heard a few people(Employees of some big firms and startups) saying their company is moving from AWS to GCP.

Enterprises have tried GCP since it claims to give more bells and whistles in terms of analytics, but for B2C AWS is the gold standard.

 

[hit&run]

Like @LurkerBaba said don't have hopes on sarkari people.

We need serious people. This is what Prof Abhinav Prakash as said. This what Rajiv Malhotra is saying. There is no ‘mai ka laal’ in India who can sit down and stitch the whole eco system together.

Few tweets from Rajiv Malhotra’sTwitter feed.




Why only I should be depressed? Come join the gloom room.

 

[hit&run]

Ok! it is 50 Billion and 10 years (Backed by AI).

This includes OS and Chip manufacturing.

After first 25 billion the ecosystem will start generating a very small percentage revenues.

Now where is the technology and R&D?

Steal, poach, buy, beg and borrow??

 

[Nicky G]

Like @LurkerBaba said don't have hopes on sarkari people.

We need serious people. This what Prof Abhinav Prakash as said. This what Rajiv Malhotra is saying. There is no ‘mai ka laal’ in India who can sit down and stitch the whole eco system together.

Few tweets from Rajiv Malhotra’sTwitter feed.




Why only I should be depressed? Come join the gloom room.

It's too big for any individual.

You are taking about essentially replicating the work of atleast 10 gigantic (100B+) companies across different aspects of technology.

The program you'd need here would make the Manhattan program look like a high school project. The only plus side, a lot of technology might exist already, if you could find a way around patents. If not, you need the entire R&D.

 

[LurkerBaba]

In terms of men hours and funds how long and what kind of investment we be looking at to create this software?

Billions of dollars probably.

But, like I said we have the capability. Flipkart already does most of this inhouse (not at AWS quality). But there's no incentive for them to provide this to external customers.

Rajiv Malhotra is actually wrong here. IT services has served as the seed of the next product based revolution in India. Its really not a problem of capability but a problem of scale. What we need are 3-4 (private) companies working at or higher than Flipkart scale. At those levels creating your own infrastructure is cheaper than buying off the shelf stuff from AWS.

What else works at Flipkart scale ? Social networks for one

 

[hit&run]

It's too big for any individual.

You are taking about essentially replicating the work of atleast 10 gigantic (100B+) companies across different aspects of technology.

The program you'd need here would make the Manhattan program look like a high school project. The only plus side, a lot of technology might exist already, if you could find a way around patents. If not, you need the entire R&D.

I think it can be done. We have to start from somewhere and fast.

 

[shade]

I think it can be done. We have to start from somewhere and fast.

Baby steps.
First one is to boot twitter and concurrently have some regulatory/legal framework dealing with SMs.

 

[Illusive]

I think it can be done. We have to start from somewhere and fast.

The risk being its a long term Investment, Indian tech companies need to pool in for this with govt backed support. It is strategic in nature and holds key to India's employment security and economic success. In short, it needs to be done regardless.

 

[Okabe Rintarou]

@shade Has answered most of it. But the data center piece needs expansion -

Data centers while important aren't enough. To build a high scale app you need ready made services and infra running on top of hardware. Only a handful of providers have such wide capabilities - AWS (Amazon), Azure (Microsoft), GCP (Google) and probably Alibaba Cloud. Basically its the software and not (just) the hardware which makes them (big tech platforms) so useful. This is why Parler couldn't simply "get off Amazon"

Don't expect GoI or the sarkari employees at NIC to make something even remotely comparable. The closest in India is probably Flipkart's Cloud Platform - Flipkart doesn't use AWS and does most things inhouse (despite funding by Microsoft). Maybe they can convert it into a desi AWS. But there's no incentive for it. Yet

I am a bit of noob in networking and IT related stuff, but one Indian alternative to AWS and Azure could be Meghdoot built by CDAC.

Cloud Infrastructure

 

[Spindrift]

It's too big for any individual.

You are taking about essentially replicating the work of atleast 10 gigantic (100B+) companies across different aspects of technology.

The program you'd need here would make the Manhattan program look like a high school project. The only plus side, a lot of technology might exist already, if you could find a way around patents. If not, you need the entire R&D.

Not really, the problem is not with investing in IT infrastructure. The Indian IT industries can easily do it, but the problem is with the pay that is offered. I know one of the principal scientist in my company that does AI related stuff i.e. developing algorithms and his CTC is a bit over 1 cr per annum (including stocks) and he was a recently offered a job by in Indian company for a CTC of 40 to 50 Lakhs per annum and when he told them his CTC they laughed and told him that the higher ups don't even make that much.

 

[tarunraju]

@shade Has answered most of it. But the data center piece needs expansion -

Data centers while important aren't enough. To build a high scale app you need ready made services and infra running on top of hardware. Only a handful of providers have such wide capabilities - AWS (Amazon), Azure (Microsoft), GCP (Google) and probably Alibaba Cloud. Basically its the software and not (just) the hardware which makes them (big tech platforms) so useful. This is why Parler couldn't simply "get off Amazon"

Don't expect GoI or the sarkari employees at NIC to make something even remotely comparable. The closest in India is probably Flipkart's Cloud Platform - Flipkart doesn't use AWS and does most things inhouse (despite funding by Microsoft). Maybe they can convert it into a desi AWS. But there's no incentive for it. Yet

The key to a desi AWS competitor is still state participation.

AWS/Google/Azure have sweetheart deals with hardware manufacturers like Intel, AMD, Supermicro, [insert datacenter hardware company], buy in astonishing volumes through single-point purchases, and overcome the high import duty/GST we levy on these things.

Single-point volume purchasing is key here. Sitting in california, someone like google is able to purchase, say, 50,000 processors from Intel, and ship them as already-purchased company assets to datacenters across the world (including India). A private desi cloud provider has to buy the same processors from Intel/AMD at a worse price (because he's buying lower volume), and go through the entire song and dance of import duties, tariffs, taxation, etc.

Alibaba and Baidu succeeded because they're basically chicom state companies who initially absorbed the costs of hardware until they could grow to a scale where they could bargain with manufacturers for volume purchases.

Riding on investor money, Jio Platforms is soon attaining this critical mass.

 

[shade]

I am a bit of noob in networking and IT related stuff, but one Indian alternative to AWS and Azure could be Meghdoot built by CDAC.

Cloud Infrastructure

They even have their own Debian Linux based operating system called BOSS.
Ofc nobody uses it.

https://bosslinux.in/boss8

 

[Nicky G]

Not really, the problem is not with investing in IT infrastructure. The Indian IT industries can easily do it, but the problem is with the pay that is offered. I know one of the principal scientist in my company that does AI related stuff i.e. developing algorithms and his CTC is a bit over 1 cr per annum (including stocks) and he was a recently offered a job by in Indian company for a CTC of 40 to 50 Lakhs per annum and when he told them his CTC they laughed and told him that the higher ups don't even make that much.

Software is easy, well easier anyway. Even there for getting quality you need investment and time.

Take a generic linux and see what it would take to get it to the level of iOS. If it's not good enough, why would peole switch?

If you want to be completely independent, the real problem is hardware.

 

[LurkerBaba]

I am a bit of noob in networking and IT related stuff, but one Indian alternative to AWS and Azure could be Meghdoot built by CDAC.

Cloud Infrastructure

To put it mildly: ROFL

Its just using OpenStack. As I said before, you can't rely on sarkaar for software. Instead govt should have the right incentives to create 3-4 private tech companies running at high scale. That'll automatically get you such infra.

Blackpill time: what is the government actually doing ? The opposite. Making it harder for startups to raise money. It fucked up Flipkart - they had to register in Singapore because of tax laws. Then they tried (and is still actively trying) to stifle ecommerce because of trader lobby. End result Flipkart was bought by Walmart and now Amazon is almost at par with it

 

[shade]

Software is easy, well easier anyway. Even there for getting quality you need investment and time.

If you want to be completely independent, the real problem is hardware.

Even Chinks are stuck here, just like jet engines, they don't have an x86 competitor, they cannot manufacture DRAM that can compete with the incumbents.

 

[Cheran]

They even have their own Debian Linux based operating system called BOSS.
Ofc nobody uses it.

https://bosslinux.in/boss8

Sometime back i tried to install this into my Oracle Virtual box, just for fun, ran into some error, then did not proceed. I might try with the newer version now. Just for fun, i have installed around 15 OS (including linux flavors, old windows including 3.1, 95, 98 etc..)