I have no idea what the laws are there, but the companies they show up at don't even necessarily have any license with Adobe. I can't imagine that is legal in India, but apparently it happens unabetted.
notorious for using disgruntled former employees for starting audits. (they did to my friends company a decade ago and found nothing..)
lots of major companies: (apple/adobe....)
Here is the Portuguese variant of such police department.
But yes, they have auditors to enforce this stuff.
IMO it's not that the engineering group is trying to not be trustworthy. It's that they have a job to do and IBM knows that this is their goal. So if we let engineering do their job, uninhibited, then there's a high chance that if the product is useful to do that they'll keep expanding it's reach internally (land and expand). This is just a slightly different spin on how the vendor allows the end user client to use their product. It's very likely the agreement between IBM and the enterprise client clearly spells this out. However, since the client is bound to pay for overages in arrears they are now at the mercy of not having as much leverage in price negotiations. A smart customer will predefine usage level bands as it pertains to discounts. For example if we're using 10k seats of RHEL we get 10% off list, 20k seats @ 20%, etc.
The disconnect is that procurement is likely disconnected from engineering for the better part between audits and they're not well prepared to have the after-the-fact negotiation to true up licensing. IBM, and others, bank on this by selling the value the customer is receiving from said products in use. The disconnect here is that had IBM enforced hard license limits the organization may say that the product is too expensive and look for an alternative because the customer is forced to buy more licensing mid-budget cycle. It's harder to rip out or change 20k installations of RHEL than it is 10k. Vendors know this and leverage this consequence of using this approach to keep the customer on the ball and chain.
So we have an enterprise that doesn't plan or communicate with itself very well and a vendor who puts out the all you can eat buffet to the hungry engineers. It's not technically wrong, and if an enterprise gets burned by it technically it's their own fault. Is it by design? For sure and my guess is that this type of model pays dividends all while being on the up and up.
Also, things can go wrong with automated license enforcement, and at large enterprise scale, that would be costly to the customer (and therefore IBM as they'd be summarily replaced). I can already imagine the HN headlines: "Flights grounded due to BGP mishap knocking IBMs license servers offline", "HDD failure costs $CORP $8.6M/day after IBM DRM prevents installation of replacement cluster"
I don't know of any large scale products that don't approach it like this today simply because the vendor doesn't want to be responsible. I'm sure some smaller products might do this, that is until they lose a customer over one of those bad situations where a customer is trying to renew a contract and they get cut off before the PO clears the backend on the vendor side.
But agreed, automated licensing does have it's fair share of pitfalls. Then again, it's like taxes. You should know, at any point in time, where you stand over the course of a year as it pertains to your over/under. The old school manual audits seem to be kept in place to keep the overages out of sight.
After seeing how clunky some of the tooling is (sqlplus is downright byzantine for no good reason), I have to wonder...
- some components of their EPM-software-suite have miserable error/warning/info messages (cryptic, misleading, etc) or even no message at all (making you erroneously think that doing something went well even if it did not).
- opening a "service request" (a support ticket) with Oracle when you have a real technical problem that needs active digging by them is often something that might make you regret having chosen IT as your work area.
Compared to that, once I'm done working and I go back home and start using Linux apps and/or submitting questions&bugs on github/linuxquestions/stackoverflow/etc and getting relatively good replies within minutes I get the feeling that I'm living in the year 3000 :)
Read and despair.
tl;dr: They dig around to see you’re paying the correct license fees.
They’re common and expensive enough that law firms offer special counseling explicitly for that.
It paid off because we got a discount on bulk licenses and met auditor inspections. I was a PC Specialist back then.
Ok thanks for posting. I will never use any RedHat product ever again.
This is the company who popularised "This page intentionally left blank".
Obtuseness is very much part of their toolkit.
That was an objectively good move though.
Graphical processor(ing?) unit. How many (cuda) cores are we talking about? /s
1. The audits can take considerable time and resources to complete (effectively, an extra cost)
2. The audits can be wrong, which sounds like a nightmare to deal with ()
As such, avoiding Redhat even if you are "paying for the software you're using" may still seem like a good idea.
 Answer in the FAQ about incorrect audits
> First, it is vital to carefully review the audit findings in detail. IBM's customers often do not receive all the license credit they are entitlted to, and IBM typically will agree to modify its initial demands if there is a mistake in the calucations. Sometimes, IBM’s compliance teams often do not have a complete file of all license agreements that an audited company may have signed. In many cases, the negotiated terms of a license agreement can include alternative counting rules or other variables that change the way license requirements are determined for certain products. If the auditors are unaware of those terms, then the calculated audit discrepancies likely will be inflated and erroneous.
There is no profit to be gained from seizing Redhat's sales operations that I can think of.
It's not even like Ubuntu, where you can quickly jump ship to Debian.
In the HN/cloud echo chamber, that might be worth pointing out.
They're the tools used by the people backing your Kube cluster :P
IPA/IdM is the gold standard for user management on a fleet of servers.
Gluster is the best replicated filesystem hands down and Ceph is the only alternative to expensive SANs that scales well.
Ansible is by far the most popular config management tool.
Foreman+Candlepin/Satellite is one of the few sane ways to do patch management.
There are a lot more things RH has under their belt that I could talk about but I just did the list that was mentioned.
Note that once you use containers, you decouple your infra tools from the app dev tool.
You can kick the can really far down the road but whether is your ops team, your VPS provider, or a massive cloud someone has to pick it up and they will likely do so using tools developed by Red Hat.
Maybe there’s a market for intra container licenses but it’s a harder sell as even there the goal is to go slim.
And that's why people DO pay for red-hat. Peace of mind, whether supported by evidence or not.
However, one the state of the art architecture is micro services, the optimal runtime is kuberentes. Once everyone is on kubernetes hardware will not matter.
Moreover, I would love a cloud based on second hand hardware (which should be much cheaper than 30%).
I’ve wanted to do this for some time, but there’s a lot of cost involved in the space (square footage itself, electricity, cooling) that you can’t cut just by running older stuff. Bandwidth and storage cost the same regardless as well (don’t buy used disks for anything remotely important).
Before the Westmere/Nehalam chips it was a different story, but anything newer is going to sip power and perform well as long as you know you’re not getting “all the perf out of it”.
It’s entirely possible to make a business out of this, but you rely on a customer base that’s inherently cheap and fussy by nature. Not all of them are, but most looking “for a deal” are cheapskates who will milk you dry in time/labor if you let them and you don’t have the margins to deal with that bullshit.
You CAN use Alpine as the hypervisor, but IBM/RedHat/Oracle customers typically don't.
And beyond kubernetes you have function as a service.
They managed to piss everyone off about 6 months after inserting IBM managers.
Typical "You're a startup. We're IBM. Our way has stood the test of time, so we'll help you out by implementing it here" stuff.
And I don't argue IBM has learned a lot of incredible business and management lessons over the years.
But you can't give a perfect playbook to someone of average intelligence, tell them they're the smartest person in the room because of the playbook, then turn them loose with other smart people and expect good things to happen.
It's questionable whether this can save them this late in the game, but there's solid seasoning behind it. IBM had few other ways to remain relevant as even crufty enterprise customers will eventually start move forward with containerization of workloads.
Low federal funds rates allow companies to acquire huge amounts of debt very cheaply.
Huge amounts of cheap debt allow companies to buy lots of stock, driving up prices.
Stock prices inflated in such a way are not supported by fundamentals and so the downside risk greatly increases.
Once the downside eventually materializes, markets drop violently. At first, the FED ignores this, but eventually it bails. The funds rate is once again lowered, so the game can continue.
In any event, the game must continue, because companies need new debt to service old debt. If the new debt was more expensive, the companies would eventually risk defaulting.
To understand the risk of corporate defaults, one must look at the importance of corporate bonds in pension funds. Ironically, the low federal funds rate is part of what drives pension funds to purchase more risky corporate debt, in order to meet their yield requirements.
All of this causes massive asset price inflation. Stock prices are detached from actual revenue, real estate prices are detached from rent income. The rich are getting richer - at least on paper - because they own most of the assets.
The CPI doesn't immediately reflect this kind of inflation, so the FED gets to claim "there is no inflation" and everything is "just fine". Well, it's not fine and they know it, they just can't really do anything about it.
Yes, but rent didn't increase the same way as property prices, which is what makes those properties overvalued in the first place.
Also, averages are misleading. Of course people will first buy property in "good" areas, not "average" areas. If you look at rent or property price development in San Francisco, you will notice that.
> As for stocks, the S&P 500 P/E ratio is about 20 which is pretty much average for recent times 
The CAPE is at 30, double the historic average and higher than at the top of the bubble in 1929 and every year since then up until the dotcom bubble. Again, averages are misleading. All you can say is that we haven't reached dotcom levels of valuation yet.
The markets have become increasingly global to the point that there is never a shortage of investors or investment capital. However, profitable, low or no risk investment opportunities do not grow on trees nearly as plentifully. Due to an overabundant supply of investors with a high demand to seek profitable low or no risk investment opportunities to park their wealth for growth via interest, and a much lower supply of profitable low or no risk investment opportunities—the interest rates to park your wealth must fall. And they have fallen quite steadily and predictably since their peak in the 1970s and 1980s... which nicely corresponds with the technological globalization and financialization of markets and marks their triumph over inflation (why inflation has not returned).
There is another reason for inflated asset prices besides cheap interest rates. Many assets are also priced to interest rates in an inverted fashion (for bonds, that is exactly the formula how they work and are priced). Just before the recession, stock buy-backs were no longer prohibited and since the great recession and the ultra low interest rate environment coupled with stock buy-backs, stocks have become a slightly more risky alternative to bonds and stock prices became just as inflated as bonds and have more or less stayed that way. You see, when you have a combination of all those factors coupled with the ability to buy back your stock, stocks are now able to compete almost directly with bonds for investors and stock prices inflate accordingly.
Because globalization, technology, and financialization happen much faster than markets and regulation can respond, and there are vested interests by investors (corporations, banks, Wall Street, etc.) to maintain the status quo via lobbying the politicians and regulators... the situation is unlikely to resolve itself via some self-regulating efficient market hypothesis.
Or, lowering interest rates creates overabundant demand for low/no risk investment opportunities...
while I agree with your analysis, it's only valid for 1 side of the coin -
this all also 'nicely corresponds' with the rise of supply-side / trickle-down economics as well..
as for inflation: is not a growth in wage disparity just as good of a measure of relative inflation over time as some small and continually modified basket of consumer goods?
Lower relative interest rates creating the overabundant demand doesn’t seem to carry as much explanatory power when you consider that interest rates were also relatively lower before the hyper-inflationary period. What changed is not just the interest rates being lower. True lower rates have some effect, it’s just not enough when you consider that the investments will also be lower growth.
Supply side economics has a hand to play here, already mentioned in deregulation, lobbying by corporate interests, decreasing tax-rates on corporations and the wealthy, financial sophistication and technology of said parties to avoid taxation and capture more of the economic progress. I guess the issue is with supply side... that it is very much USA centrist and is not entirely explanatory globally. The USA does have the world’s reserve currency though. But take Japan for instance, it started the trend of low interest rates and QE forever because of globalization issues with currencies and economies, not supply side economics.
I think supply side movement isn’t really incompatible with my explanation and it actually fits inside it instead of my explanation fitting inside supply side changes. I think it is helpful to remember that the neoliberal economic and libertarian agendas are responsible for all of this in any event.
Income inequality is another issue... economic progress for the laborers being incremental at best while the progress of capital owners has been explosive. There is a tie-in possibly, that the income share of labor is being hallowed (or crowded) out by capital. That the economy is making laborer and consumer activity highly irrelevant in the grand scheme of things. The economy is more concerned now about retaining capital that be can pulled out en mass on a whim and major investments can be postponed on an indefinite basis for years until the conditions are capital-perfect again.
The relevance of consumption is diminishing and laborers and consumers are increasingly living in an economy that is not concerned with balancing the affordability of housing and education and health care with how the rest of the economy is performing... especially the items that are rather inelastic and you require to live or progress.
Well, zero is kind of a special number. Anything significantly different from zero should create a very different dynamic (with a hyperbolic reaction of the market the closer you get to zero).
That said, the stock market is never "supported by fundamentals" like the GP wants. Also, just because the interest rate is low, it does not mean that all companies are under water. Besides, I don't think anybody already knows the consequences of zero interest rates well enough to say with certainty that it's bad, but it's different from non-zero.
If I could conceive of a way of calculating the optimal federal funds rate, I'd be owed a nobel prize in both economics and physics. Even then, the Fed probably wouldn't listen to my advice.
The secular decline in real interest rates over almost 40 years is global phenomenon. Increasing demand for safe and liquid assets and decline in world economic growth might be factors.
In the US the amount of money in mutual funds, hedge funds, ETFs, insurance companies, 401Ks etc has exploded while supply has decreased. The number of listed stocks in US markets has dropped dramatically. This is genuine change in supply and demand.
That's a fair point and that wasn't my intention, but you can only fit so much context into a comment. I'm not even saying they're "doing it wrong", but that they are in a pickle. Of course interest rates are low globally and it's not just domestic capital bidding up prices.
> In the US the amount of money in mutual funds, hedge funds, ETFs, insurance companies, 401Ks etc has exploded while supply has decreased.
Well alright, but why has it exploded? Why do all these allocations into riskier asset classes take place? Is it all just greed, or is it the interest rate environment?
> Huge amounts of cheap debt allow companies to buy lots of stock, driving up prices.
> Stock prices inflated in such a way are not supported by fundamentals and so the downside risk greatly increases.
The price is driven up by "demand" but the demand is not "real" because it wasn't due to the business actually _innovating_ but actually just "stock changing hands"?
> Once the downside eventually materializes, markets drop violently. At first, the FED ignores this, but eventually it bails. The funds rate is once again lowered, so the game can continue.
When/how does that happen?
> All of this causes massive asset price inflation.
There is no way for it to "not" right? It's sort of like measuring volume of action, but the action didn't actually produce any real "value"?
> The CPI doesn't immediately reflect this kind of inflation, so the FED gets to claim "there is no inflation" and everything is "just fine". Well, it's not fine and they know it, they just can't really do anything about it.
Is there a better measure that _does_ reflect it? It sounds like it can be modeled with corporate debt / buyback on a graph?
Again, layperson here, hoping to learn stuff.
The demand is real, but the fundamentals do not support it. Fundamentals include things like price vs. earnings or debt vs. assets, dividend yields, etc.
If you like at this graph of price vs. earnings, you can can see dramatic rises relating to stock market bubbles:
> When/how does that happen?
It has already happened and we're in the middle of it. When the markets tanked in late 2018, the Fed stopped hiking. When the markets tanked again in May, the Fed announced a rate cut. Now that the rate cut is in place, the markets are tanking again, perhaps because a bigger rate cut was priced in. We'll see what happens in the coming months.
> There is no way for it to "not" right? It's sort of like measuring volume of action, but the action didn't actually produce any real "value"?
I guess that's a fair way of looking at it.
> Is there a better measure that _does_ reflect it?
Of course there are indicators that you can use to show that this kind of inflation is taking place, like the S&P 500 PE ratio I showed earlier, or rent income vs. property prices.
Taken to the extreme, there would finally be 1 share left held by someone and thus that person would own 100% of the company and be entitled to 100% of the profits from said company. The company cannot just buy the last stock because the stock itself is worth all the cash the company holds and the future profits it will make. Saving up more cash to buy the final stock just raises the price because the stock is entitled to that cash and nobody is going to loan a company enough cash to be worth the expected future profits of said company which makes up the remainder of the shares value.
Funnily enough though, I think they would still do share buybacks by first splitting the stock and then buying a subset of the split stock because then the sole owner would only have to pay taxes on the sale price minus the price he paid for the stock whereas dividends are taxed in their entirety.
The last share is worth whatever the person that holds it is willing to sell it for, right? A share of IBM is worth $136 right now because there are two parties willing to buy/sell a share of IBM at that price.
I realize this would never happen in a million years, but could a company exist that is owned by nobody?
It’s not that unusual to have entities that essentially own themselves, in the sense that nobody is entitled to the profits and the board or managers elect their own successors. This is the case for foundations and trusts in many countries, usually with some kind of charitable purpose. Some of them even operate businesses. It might be possible for a company to take on some debt, buy out all its shareholders, and reorganize into a charity.
One example is Dell:
Also see Elon Musk's infamous tweet about Tesla:
Interestingly, the company didn’t use buybacks to become private but it did use buybacks to become public again last year . It was done by buying back some piece of its business that remained traded in the stock exchange following the acquisition of EMC when Dell was private .
But theoretically if a company do keep buying its own shares it can reduce their number to the point where the company is no longer required to be “public” and the shares can be delisted. However the important number is how many shareholders the company has, not how many shares (but obviously a company with 300 shares cannot have more than 300 shareholders).
I don’t know if that has ever happened. Reducing the numbers of shares (and shareholders) can also be done be reverse splitting. For example a company that is doing so badly that the share price goes to zero and regularly do inverse splits (making one $5 share from 100 $0.05 shares) will end up with a handful of shares. Or a company which is not failing may try to get rid of minority shareholders by doing reverse splits (if you don’t have enough shares to get the new one you get cash instead).
The usual way to reduce the number of shareholders is for someone to make a tender offer. And in that case there are rules about what happens with those who didn’t choose to sell. That’s what happened with Dell, that another comment mentioned, which was bought by Silver Lake.
>Huge amounts of cheap debt allow companies to buy lots of stock, driving up prices.
Companies don't use debt to buy stock. Taking on the debt to do that would depress the stock price the same amount that the buying would attempt to appreciate it.
>All of this causes massive asset price inflation.
This would be reflected in the CPI.
>Stock prices are detached from actual revenue
No they aren't. See UBER/LYFT/AAPL.
>so the FED gets to claim "there is no inflation"
The Fed does not claim that. They claim inflation is nominal near their intended target.
Finally, stop saying FED. It's not an acronym.
Sure they do.
By the way, the claim “Taking on the debt to do that would depress the stock price“ doesn’t seem correct. The influence of leverage on valuation is complex and depends on the cost of debt and equity, tax shielding considerations, the increase in risk due to the financial leverage... and essentially on what the money is going to be used for.
It often makes more sense to use debt than cash for buybacks.
If you make that claim, you should do some actual debunking instead of just nitpicking.
> Companies don't use debt to buy stock.
Yes, they do.
> Taking on the debt to do that would depress the stock price the same amount that the buying would attempt to appreciate it.
Maybe if valuation was strictly tied to fundamentals, which it isn't.
> This would be reflected in the CPI.
Eventually it will be, but that can take a while, because many of the prices involved in the CPI are sticky.
> No they aren't. See UBER/LYFT/AAPL.
I'm sorry, I meant to write "profits" there. Sure, UBER and LYFT may have a lot of revenue, but they're losing a lot of money in the process. I wouldn't say Apple is overvalued, but clearly the market is taking its lack of growth prospects unkindly.
> The Fed does not claim that. They claim inflation is nominal near their intended target.
Fair enough, to be precise, that is what they are saying, and what I mean by "no inflation (to worry about)".
> Finally, stop saying FED. It's not an acronym.
I'm aware of that, but I think it looks better this way. If writing it that way annoys just one nitpicker on the internet, it's worth it.
Banks do the same thing. Inflation is like a cone of molasses that is being pored on the banks. It then slowly expands. So banks get the cheap money first and lend it to people that haven’t been hit with inflation yet.
If they do this fast enough, they make money off arbitraging the inflation rates. Of course, the CPI doesn’t capture any of this.
But I also think this confuses topics. You are just talking straight up arbitrage that has nothing to do with being a bank. That arbitrage from high supply area to low suppy area will lower the price. The arbitrager makes some money on the transactions, but the net effect is to push the price down. Not that great for the huge socks on loans already on the banks balance sheets.
While the way the Fed operates, currency enters our economy through bank deposits turned into loans (not as true as it once was in the post QE world) those loans are going to be repaid 5+ years when inflationary effects have already started pushing the dollars down that they are being paid back in. Plus, the point of the extra reserves is to drive down rates since there will be a greater supply. They get hosed in both ways.
Moreover, being a bank is about arbitrage. Traditionally, banks arbitrage between short term loans (the federal funds rate and the overnight rates) and long term debt. i.e Banks borrow short-term and lend long-term. The way they make money is taking on risk: converting a lot of risk (long-term loans) into short-term risk (the banks borrowing at the overnight rate). This is where quants and models come in. It takes a lot of skill to turn that long-term risk into short-term risk.
> Stock prices inflated in such a way are not supported by fundamentals
Inflation - what your describing, more dollars chasing same production - is a fundamental factor and doesn't affect rate of return in the time scales you are talking about.
> All of this causes massive asset price inflation. Stock prices are detached from actual revenue,
No, then there would be an arbitrage opportunity (if you so easily see it obviously the smart, deep pocketed money would see it too). Stock prices are still set by future income streams but now that is nominally pumped by inflation so they are bid up. Since equity prices are forward looking their prices rise faster than the backwards looking cpi.
There really isn't a lot of extra dollars sloshing around though. All the forward looking indexes of inflation are down: commodity crb index, metals index (including gold), tips spreads - all down in the last six months.
The federal funds rate greatly affects credit rates, that's why it's a primary instrument in financial policy.
> Pensions and retirement funds are only affected by the overnight rates in so much they use hold short term reserves as cash equivalents.
That's the direct effect, but because all other credit rates are affected, so are corporate bonds, which are a major component of pension fund allocations.
> All the forward looking indexes of inflation are down: commodity crb index, metals index (including gold), tips spreads - all down in the last six months.
> No, then there would be an arbitrage opportunity (if you so easily see it obviously the smart, deep pocketed money would see it too).
Why? Who says that prices aren't going to stay "permanently inflated"? I'm talking about downside risk, not a guaranteed and easily timed market selloff.
Sure, if you look at commodities, there is no inflation. If you look at the CPI, there is no inflation. It depends on where you look. Gold is way up though, regardless of what the index says.
But the fed doesn't control long term rates (probably wish it did). The short end is heavily driven by technical factors including the overnight rate (even then that is only a target and it doesn't anyways follow what the fed wished it was). The long end isn't the same. It is driven by return on capital (higher returns both seek out loans and can afford to pay more for them).
If there was long term mispricing we would see it in the tips market, and we don't. The research shows the opposite - been little arbitrage exists in the market and it is extremely well allocated.
The Fed only signaled a return to rate cuts two months ago.
> But the fed doesn't control long term rates (probably wish it did).
I never said they "control" long-term rates, but they greatly influence them, especially in terms of how low they can go.
> The long end isn't the same. It is driven by return on capital (higher returns both seek out loans and can afford to pay more for them).
> If there was long term mispricing we would see it in the tips market, and we don't.
I don't see that. It takes quite a while for overvalued asset prices to make a dent in the CPI, by that time it's more likely that the bubble has popped.
That said the picture painted by this article is that doing nothing wasn't an option, declining revenues is not a good picture at a time when the competition are posting record numbers...
These are not good projections. It's really hard to understand how RedHat acquisition makes any financial sense whatsoever if this is the actual effect. RH annual revenues were $3.36B for FY2019.  That's a tiny fraction of IBM's current revenue.
(Whether that justifies the valuation I don't know enough to guess on.)
And if they can sell some of them RHEL on POWER9 they will be very happy to take that money as well :)
Given the market is moving, even IBM's major customers would eventually migrate to newer technologies. (On the order of decades)
By providing something that mostly satisfies that need, IBM essentially drilled a relief well to take the pressure off. When they're ready to migrate, they could get 80% value of what a non-IBM migration would get them by moving to IBM's product.
For a lot of large, slow, conservative orgs, that's good enough.
Yes, but that’s only a coincidence. What they have really been doing is transitioning from proprietary to commodity hardware. Which IBM no longer makes. Where does IBM make money? Mainframe sales. There can be only one winner here... IBM will need to kill off their mainframe business to justify this purchase, something they have proved very reluctant to do.
IBM’s mainframe margins are safe for decades to come.
But realistically mainframes are a small part of IBMs business so there are lots of reasons for IBM to buy Red Hat that are entirely unrelated to mainframes or POWER hardware
IBM is mainly a consultancy business, so they basically back integrated with the cloud os - kubernetes, or more specifically in this case - open shift.
The only issue here is that are around 23 or more certified kubernetes distors, so this might have been an expansive price to pay.
If anything I am skeptical about Kubernetes' future. Not in the short term, k8s/docker are the MongoDB of the late 2010s and the same sorts of people are jumping on the bandwagon. But look at the discussion about the AMD EPYCs in Google Cloud. Big machines are now getting so big that a bog standard OS running a bunch of processes is going to provide more than enough power for even very large websites.
One machine now provides more power than a whole rack used to, and horizontal scaling technologies are expensive. A research paper from 2013 showed that just scaling up beat the performance of scaling up for virtually all jobs analysed, which came from Microsoft, Yahoo, and Facebook. So the sorts of companies that are poster-children for this kind of cloud technology.
There's not much career or intellectual capital to be gained by just running things on big boxes, but, eventually the cost effectiveness of this will become hard to ignore.
A cloud native system is a system that manage its resources via an API. The vendor locking in this case is the cloud API. I.e. if I want a resource (e.g. a VM) , I need to use the EC2 api.
With kubernetes, I write to the kuberentes API which is open source, cloud vendor natural.
However, and this the key to my comments - kubernetes is the platform of platforms. It is an API platform and not just an API.
In the future (some would say near future), all of the cloud services will be offered as kubernetes APIs (via a set of domain specific CRD). I am currently working on the AI platform.
As for the big machines comments. I absolutely agree. However, I would like to ignore the actual hardware (banning some sort of HA that would require at least 2 machines).
If (big if) some standardized variant of kubernetes "wins."
K8s was transparently Google's strategic shot at forcibly commodifying AWS.
But everyone has their own agenda. Google, Microsoft, and (to a lesser extent) Amazon could survive in a world of fully commodified compute. Give their compute assets are essentially surpluses from running their other businesses (advertising, software sales, and online retail).
So none of them have a strategic imperative to derail k8 evolution.
But managed, specialized cloud services are lucrative to all of them (especially Amazon). So their best-case isn't a fully commodified future.
This is mythology. It is mostly not actually true.
I doubt very much the first deployment of Gmail looked anything like its footprint today.
More regions doesn't necessarily stem solely from customer demand.
I expect the vast majority does, but without numbers, that's just a guess.
An AI service is much more than submitting a CSV file to a cloud and getting a model back, or even a web end point.
1) watch for changes in the input data.
2) watch for model decay.
3) watch for latency.
4) retrain (preferably automatically).
5) redesign your pipeline and imputation methods when the data changes.
6) watch for problems in the input data during serving
7) Scale out or Scale down.
8) Be optimal with the training resources (for example, train on spot instances).
9) Canary and shadow release of model
10) Data and model versioning.
Not to mention privacy and price.
What happen today is that there are no offering of alternative to the cloud services, so there is no point of reference.
Let's stipulate that most analytic applications can run on single servers. You still need a way to schedule work efficiently when the applications in aggregate require more than a single host. Kubernetes does that very well.
That's leaving out considerations like HA. The paper is notably silent on this topic. To keep systems up across failures as well as maintenance you need replicas in different locations. There's an implication in the paper that everyone is just going to run off S3, which is assumed to be globally available. Not everyone makes that choice for a variety of reasons. So we will continue to see shared nothing designs that require placement across hosts, racks, AZs, etc. Kubernetes is efficient at deploying these.
Kubernetes doesnt solve the vendor lock in issue completely. You still need workers, external dns, databases, gateways and other to operate.
The strength of Kubernetes comes in development speed. And thats pretty much it. You can push something out pretty fast and scale it. Ah, this and the fact its easier to find newbs familiar with k8s (they often have no idea how to deal with bare-metal).
I wouldnt say the amount of operational work vs normal cloud management is simplified. Managing kubernetes cluster is painful, Ive witnessed 3 updates of EKS where each one broke some part of the cluster: dns / autoscaling / monitoring etc.
But hey it is fast. Building runtime image with immutable layers is a huge adventage in itself.
Also see https://github.com/kubernetes-incubator/external-dns.
The better question is, why did it cost so much? First of all, Red Hat makes a lot of money and it has a solid market position.
Secondly, you tend to overpay in an acquisition, because it is a huge amount of sudden demand.
Lastly, pretty much all stock is "overvalued" right now, because money is cheap (see my other comment).
> Why would IBM put such a big bet?
Because it is a giant corporate monstrosity that cannot really innovate by itself anymore. It still can buy things though.
Seems like IBM is hedging their future on Linux.
Microsoft? Tencent? Dell? gulp Oracle?
I understood that it's often used in corporations as an alternative for RedHat => maybe Suse will have to merge as well with some other company if the combo RedHat&IBM in/directly shrink their revenues?
E.g. maybe Suse will be excluded from IBM's support? ( https://www.suse.com/partners/alliance/ibm/ )
It's perhaps the single biggest "alternative" strategy ETF.
A lot of quant market neutral hedge funds only return 3-5% per year. Compared to the S&P, this looks pretty bad. But the goal of these funds isn’t to provide the best returns, it is to provide the best risk adjusted returns. You may ask, why would I care about that?
The answer is that you can synthetically match any return by either buying or selling (borrowing) at the risk free rate. So if I had a fund that had returns of 4% and volatility of 2%, I could turn it into 2 and 1, 8 and 2, or 16 and 4, or whatever.
Historical yield is not everything. Things change, past performance is not necessarily a predictor of future success.
Completely conflates Red Hat with RHEL and makes it sound like IBM purchased a linux distro for 34b.
Sounds like a first-world financial scenario, no?
The newer Thinkpads from Lenovo are more like wannabe Macbooks; I miss batteries you could swap out without opening the computer and Thinkpads you could open without using a spudger.
As a real life example, the other day I needed to migrate /var on a Linux box. The freely available CentOs guides covered it well, but when I was done the applications were having a lot of issues. Going through Redhat’s documentation it quickly became obvious that the other articles neglected to mention rebuilding the grub boot files and getting SELinux running on the new directory.
Package search is on https://git.centos.org/
For example, libvirt is at version 4.5 in CentOS 7.6, while CentOS 7.0 shipped with version 1.1:
Whereas in Ubuntu 18.04 LTS, it's at version 4.0:
The RHEL kernel's version number is doesn't signify as much as you'd think, there are whole subsystems backported into it from later upstream kernel releases.