IBM Finally Abandons Mainframe Division After 100 Years Of Operations

[duomaxattacks]

June 19 2019 IBM completed its acquisition of Red Hat. Nearly a year and a half later the company has so much confidence in RedHat's parent devision, The Cloud Devision, that they no longer want to be associated with Mainframes and other Infistructure services. The Cloud Devision is the only growing devision inside IBM. Due to massive cost cuts in R&D, labor, and investment as well as using creative accounting practices IBM was able to delay some of the huge issues its been facing since the 90's at bay until around 5 - 6years ago. The traditional mainframe devision has been threatened by basic x86 servers for decades, and the last decade of evolving cloud hosting companies has effectively nailed closed most of IBM into a weighted coffin and tossed them into the ocean. IBM has been a long time investor in Linux, but also a sort of fair weather friend in regards to percentage of spend reletave to size of company. Remember in 2009 IBM was the forth largest company in the world. And its Linux investments have played critical roles in the early years. IBM invested a billion dollars in 2000 on Linux, but has since been easily overshadowed by others Including Facebook who has spent an estimated 16billion dollars on open source technologies. It's presumed that IBM both hopes to jump the stock price by changing their narrative, and change the management focus of the company and investors into greater R&D spend. Other more successfull Coud Giants like Google, Amazon, and Microsoft are heavily invested in A. I. and open source. P. S. Open source is not necessarily a silver bullet for every Cloud company, at least in my opinion. But it's hard as a private company to get free spec work if your customers and enthusiasts can't see your code.

 

Summary

 10 Years of declining infrastructure management, (aka mainframe,) services sales and IBM has decided to spin off the division into its own company.

 

Quotes

Quote

 “We believe we can deliver strong growth within the company with the financial flexibility we will create with this transaction,” CEO Arvind Krishna told CNBC’s “Squawk on the Street.”

----------------------------------------------------------------------

“IBM will focus on its open hybrid cloud platform and AI capabilities,” CEO Arvind Krishna said earlier Thursday in a prepared statement. “NewCo will have greater agility to design, run and modernize the infrastructure of the world’s most important organizations. Both companies will be on an improved growth trajectory with greater ability to partner and capture new opportunities – creating value for clients and shareholders.”

----------------------------------------------------------------------

“The success we’ve had with Red Hat gives us confidence that this is the right move,” CEO Arvind Krishna

 

Quote

 

 

My thoughts

Layoffs, outsourcing, and massive cuts to R&D for decades led to little room for new internal business growth. If they wish to maintain the growth of the Cloud Division, aka Watson and Redhat, then they need to kill off the old guard. Hopefully they stop the disorganized cost management at all costs strategy and turn around like Microsoft has under  Satya Nadella. It would suck to see RedHat, formerly the largest open source company and the birthplace of kvm*, crumble like the rest of IBM has.

 

 

 

Sources

. https://www.cnbc.com/2020/10/08/ibm-shares-surge-on-plans-to-spin-off-unit-into-separate-publicly-traded-company-.html

. https://www.theregister.com/2019/07/18/ibm_q2_fy2019_decline/

. https://www.bloomberg.com/news/articles/2020-01-21/ibm-beats-estimates-ekes-out-revenue-growth-on-cloud

. https://www.redhat.com/en/about/press-releases/ibm-closes-landmark-acquisition-red-hat-34-billion-defines-open-hybrid-cloud-future

. https://finance.yahoo.com/quote/IBM/

. https://en.m.wikipedia.org/wiki/Kernel-based_Virtual_Machine

. https://itsfoss.com/ibm-invest-1-billion-linux/

. https://www.linuxfoundation.org/blog/2020/08/facebooks-long-history-of-open-source-investments-deepens-with-platinum-level-linux-foundation-membership/

 

 

[Radium_Angel]

5 minutes ago, duomaxattacks said:

Cloud

Fuck that noise. I have a perfectly powerful system right here on my desktop, but let's squander that power and go online only!

Works great, until it doesn't.

Office 365 has suffered multiple outages over 202.

 

No thanks, I like my data where I know it's safe, in my hands.

 

Shame IBM is killing off mainframes. Inevitable I suppose. 

[Pickles von Brine]

12 minutes ago, Radium_Angel said:

Fuck that noise. I have a perfectly powerful system right here on my desktop, but let's squander that power and go online only!

Works great, until it doesn't.

Office 365 has suffered multiple outages over 202.

 

No thanks, I like my data where I know it's safe, in my hands.

 

Shame IBM is killing off mainframes. Inevitable I suppose. 

For a consumer, cloud may or may not make sense, but for a business, it makes a ton of sense. Why host your own hardware when you can just use some else's for cheaper. A few companies I know who went to cloud have saved tens of thousands a month just from not hosting the hardware anymore. 


Though i do agree, mainframes have been dying. X86 systems can be built in similar ways for cheaper. Though the redunency inside of a mainframe is something else. I know the zEnterprise/196 MCM could have a full processor failure and not notice. So, really depends on the use I guess. 

However, there are certain applications a mainframe is going to win over a server farm. Though my area of knowledge stops here. 

[Kilrah]

7 minutes ago, Pickles - Lord of the Jar said:

I know the zEnterprise/196 MCM could have a full processor failure and not notice.

It's just not necessary anymore, the paradigm has shifted redundancy from the machine level to the app level. Instead of having a super reliable computer that can have a CPU die and not notice apps are designed to run on multiple machines and not care if one dies. Similar effect, plus has other advantages like offering scaling possibilities the mainframe approach could never achieve.

[Trik'Stari]

4 hours ago, Pickles - Lord of the Jar said:

For a consumer, cloud may or may not make sense, but for a business, it makes a ton of sense. Why host your own hardware when you can just use some else's for cheaper? 

Because that cost will eventually be offset by the total number of service outages experienced by your company that prevent work from getting done, deadlines being met, and customers being satisfied to the extent that they continue to pay you.

 

Not saying Cloud services don't have a use, just saying they should NEVER be the sole option for any company.

[williamcll]

25 minutes ago, Trik'Stari said:

Because that cost will eventually be offset by the total number of service outages experienced by your company that prevent work from getting done, deadlines being met, and customers being satisfied to the extent that they continue to pay you.

 

Not saying Cloud services don't have a use, just saying they should NEVER be the sole option for any company.

Shouldn't that be the blame of the ISP and not the cloud provider?

[Trik'Stari]

1 hour ago, williamcll said:

Shouldn't that be the blame of the ISP and not the cloud provider?

Depends on who is to blame, but thank you for pointing out yet another point of failure for the cloud. So now you technically have two parties that can F up and result in your company or organization being incapable of functioning.

[Kilrah]

6 hours ago, Trik'Stari said:

Because that cost will eventually be offset by the total number of service outages experienced by your company that prevent work from getting done, deadlines being met, and customers being satisfied to the extent that they continue to pay you.

Firstly in-house outages are a thing too, and when you get enterprise level cloud storage it's a little more reliable that what consumers get. And if you develop your app to be in the cloud it's easy to develop it so as to be deployed across 2 different cloud providers.

 

5 hours ago, Trik'Stari said:

but thank you for pointing out yet another point of failure for the cloud.

You set up 2 redundant ISPs as well. 

 

 

[leadeater]

13 hours ago, Pickles - One of the Jar said:

Why host your own hardware when you can just use some else's for cheaper

LOL, no. For cheaper is often not the case, more flexible and easier to manage sure but cost comparisons can be very hard. However always keep one very important thing in mind, what ever you are purchasing or subscribing to is from a for profit business which means there is absolutely room for you to do the same thing yourself at a lower cost margin than they are charging.

 

Cloud options make the most sense in relation to cost for small business that basically don't have the resource requirement to run more than one server or even fully utilize one server, that is the most clear cut 'Cloud is cheaper' example you can get. Once you start getting in to the scale of multiple servers worth of resource requirement TCO calculations get very hard very fast.

 

For example a single database VM from either Azure or AWS costs the same for a single year as an entire HPE server costs (2x 6256 + 386GB ram) is to buy that has a 5 year life cycle and that server is many times more resources and higher performance and can host multiple database instances. Database Platform as a Service options are cheaper but also much lower performance too and the costing of them is needlessly complicated much of the time.

[leadeater]

8 hours ago, williamcll said:

Shouldn't that be the blame of the ISP and not the cloud provider?

No because cloud providers have LOTs of outages and issues, Office 365 for example is always having problems it's just that these happen in different ways and effect different things and often isolated to a single Region. I've honestly lost count of the number of Office 365 issues this year alone, in contrast to when we ran it ourselves and had zero over 5 years.

 

ISP's are not the problem, they are in fact many times more reliable.

[Trik'Stari]

2 hours ago, Kilrah said:

Firstly in-house outages are a thing too, and when you get enterprise level cloud storage it's a little more reliable that what consumers get. And if you develop your app to be in the cloud it's easy to develop it so as to be deployed across 2 different cloud providers.

 

You set up 2 redundant ISPs as well. 

 

 

1. We've had in-house outages as well. It's funny to me because it brings our work to a complete stop, and still the management refuses to accept it's because they use the cheapest hardware they can find.

 

2. "Redundant ISP's? Sounds like a waste of money, we can cut that and reduce overhead by some small percentage and I can get a bigger bonus!" - Management.

 

The year I started working for this company, we picked up a customer that brought in an additional 30,000 laptops, and corporate started complaining that we were using too many paper towels and too much windex. I shit you not, "you didn't use this many last year!". Someone finally explained it to them, they still demanded that management attempt to tell us to use only 4 paper towels per day. We were expected to repair something like 20 laptops a day. Still are, but I'm no longer in the repair room, hence the past-tense.

 

Never underestimate the capacity for stupidity by management, and never assume they have ANY amount of foresight at all.

[Dom1252]

14 hours ago, Kilrah said:

It's just not necessary anymore, the paradigm has shifted redundancy from the machine level to the app level. Instead of having a super reliable computer that can have a CPU die and not notice apps are designed to run on multiple machines and not care if one dies. Similar effect, plus has other advantages like offering scaling possibilities the mainframe approach could never achieve.

mainframes are further with this, whole systems have redundancy in virtual and in physical way, depends how you set it up 

basic approach is to have 2 copies of one system on each of 2 locations (if you want more, get more) 

that means... if one of your CPU dies, no problem... your I/O cards start dying? no problem... storage? no problem, fire on location? no problem... fire on one location, motherboard death on other and few cpu's death on top of it? no problem, still running (there might be few seconds downtime tho, depends which machines and what SW you're running) 

of course, if you throw nukes on both locations, you might be screwed... but if you need system like that, build it in bunker 

 

what scaling possibilities Z cannot achieve in businesses where it competes? 

[Kilrah]

1 hour ago, Dom1252 said:

what scaling possibilities Z cannot achieve in businesses where it competes? 

What most large scale cloud apps do - running on 100 machines, and when you get a demand spike be able to scale to 1000 machines to take the unexpected loads within minutes. Then when it falls back drop to 50 machines so as not to pay for unused hardware.

[Pickles von Brine]

3 hours ago, leadeater said:

LOL, no. For cheaper is often not the case, more flexible and easier to manage sure but cost comparisons can be very hard. However always keep one very important thing in mind, what ever you are purchasing or subscribing to is from a for profit business which means there is absolutely room for you to do the same thing yourself at a lower cost margin than they are charging.

 

Cloud options make the most sense in relation to cost for small business that basically don't have the resource requirement to run more than one server or even fully utilize one server, that is the most clear cut 'Cloud is cheaper' example you can get. Once you start getting in to the scale of multiple servers worth of resource requirement TCO calculations get very hard very fast.

 

For example a single database VM from either Azure or AWS costs the same for a single year as an entire HPE server costs (2x 6256 + 386GB ram) is to buy that has a 5 year life cycle and that server is many times more resources and higher performance and can host multiple database instances. Database Platform as a Service options are cheaper but also much lower performance too and the costing of them is needlessly complicated much of the time.

I learned something!

[Doobeedoo]

We will all become cloud. 

[DaRk0]

Well, I wouldn't say abandoned.

 

These are just managed infrastructure services which means clients that rent Z as IaaS. Also note that IBM's managed services have more offerings than just Z in their portfolio.

 

On prem IBM Z clients are staying with IBM and there are few new offerings coming to on prem clients that I can't talk about at this moment, but they might be publicly announced soon.