1 |
On 08/01/2016 01:03 PM, Rich Freeman wrote: |
2 |
> On Mon, Aug 1, 2016 at 12:49 PM, J. Roeleveld <joost@××××××××.org> wrote: |
3 |
>> On Monday, August 01, 2016 08:43:49 AM james wrote: |
4 |
>> |
5 |
>>> Sure this part is only related to |
6 |
>>> transaction processing as there was much more to the "five 9s" legacy, |
7 |
>>> but imho, that is the heart of what was the precursor to ACID property's |
8 |
>>> now so greatly espoused in SQL codes that Douglas refers to. |
9 |
>>> |
10 |
>>> Do folks concur or disagree at this point? |
11 |
>> |
12 |
>> ACID is about data integrity. The "best 2 out of 3" voting was, in my opinion, |
13 |
>> a work-around for unreliable hardware. It is based on a clever idea, but when |
14 |
>> 2 computers having the same data and logic come up with 2 different answers, I |
15 |
>> wouldn't trust either of them. |
16 |
> |
17 |
> I agree, this was a solution for hardware issues. However, hardware |
18 |
> issues can STILL happen today, so there is an argument for it. There |
19 |
> are really two ways to get to robustness: clever hardware, and clever |
20 |
> software. The old way was to do it in hardware, the newer way is to |
21 |
> do it in software (see Google with their racks of cheap motherboards). |
22 |
> I suspect software will always be the better way, but you can't just |
23 |
> write a check to get better software the way you can with hardware. |
24 |
> Doing it right with software means hiring really good people, which is |
25 |
> something a LOT of companies don't want to do (well, they think |
26 |
> they're doing it, but they're not). |
27 |
> |
28 |
> Basically I believe the concept with the mainframe was that you could |
29 |
> probably open the thing up, break one random board with a hammer, and |
30 |
> the application would still keep running just fine. IBM would then |
31 |
> magically show up the next day and replace the board without anybody |
32 |
> doing anything. All the hardware had redundancy, so you can run your |
33 |
> application for a decade or two without fear of a hardware failure. |
34 |
|
35 |
Not with todays clusters and cheap hardware. As you pointed out |
36 |
expertise (and common sense) are the quintessential qualities for staff |
37 |
and managers..... |
38 |
|
39 |
> |
40 |
> However, you pay a small fortune for all of this. |
41 |
|
42 |
Not today, that was then those absorbant prices. Sequoia made so much |
43 |
money, I pretty sure that how they ultimately became a VC firm? |
44 |
|
45 |
|
46 |
|
47 |
> The other trend as |
48 |
> I understand it in mainframes is renting your own hardware to you. |
49 |
|
50 |
Yes, find a CPA that spent 10 years or so inside the IRS and you get |
51 |
even more aggressive profitibility vectors. Some accouants move |
52 |
hardware, assest and corporations around and about the world in a shell |
53 |
game and never pay taxes, just recycling assets among billionares. It's |
54 |
pretty sickening, if you really learn the details of what goes on. |
55 |
|
56 |
> That is, you buy a box, and you can just pay to turn on extra |
57 |
> CPUs/etc. You can imagine what the margins are like for that to be |
58 |
> practical, but for non-trendy businesses that don't want to offer free |
59 |
> ice cream and pay Silicon Valley wages I guess it is an alternative to |
60 |
> building good software. |
61 |
|
62 |
Investment credits, sell/rent hardware to overseas divison, then move |
63 |
them to another country that pays you re-locate and bring a few jobs. |
64 |
Heck, event the US stats play that stupid game with recruiting |
65 |
corporations. Get and IRA career agent drunk some time and pull a few |
66 |
stories out of them..... |
67 |
|
68 |
>> You have seen how "democracies" work, right? :) |
69 |
>> The more voters involved, the longer it takes for all the votes to be counted. |
70 |
>> With a small number, it might actually still scale, but when you pass a magic |
71 |
>> number (no clue what this would be), the counting time starts to exceed any |
72 |
>> time you might have gained by adding more voters. |
73 |
>> |
74 |
>> Also, this, to me, seems to counteract the whole reason for using clusters: |
75 |
>> Have different nodes handle a different part of the problem. |
76 |
> |
77 |
> I agree. The old mainframe way of doing things isn't going to make |
78 |
> anything faster. I don't think it will necessarily make things much |
79 |
> slower as long as all the hardware is in the same box. However, if |
80 |
> you want to start doing this at a cluster scale with offsite replicas |
81 |
> I imagine the latencies would kill just about anything. That was one |
82 |
> of the arguments against the Postgres vacuum approach where replicas |
83 |
> could end up having in-use records deleted. The solutions are to |
84 |
> delay the replicas (not great), or synchronize back to the master |
85 |
> (also not great). The MySQL approach apparently lets all the replicas |
86 |
> do their own vacuuming, which does neatly solve that particular |
87 |
> problem (presumably at the cost of more work for the replicas, and of |
88 |
> course they're no longer binary replicas). |
89 |
|
90 |
Why Rich, using common sense? What's wrong with you? I thought you were |
91 |
a good corporate lacky? Bob from accounting has already presented to |
92 |
the BOD and got approval. Rich, can you be a team player (silent idiot) |
93 |
just once for the team? |
94 |
|
95 |
|
96 |
> |
97 |
>> |
98 |
>> The way Uber created the cluster is useful when having 1 node handle all the |
99 |
>> updates and multiple nodes providing read-only access while also providing |
100 |
>> failover functionality. |
101 |
> |
102 |
> I agree. I do remember listening to a Postgres talk by one of the |
103 |
> devs and while everybody's holy grail is the magical replica where you |
104 |
> just have a bunch of replicas and you do any operation on any replica |
105 |
> and everything is up to date, in reality that is almost impossible to |
106 |
> achieve with any solution. |
107 |
|
108 |
Yep NoSQL is floundering mightily when requirements are stringent and |
109 |
other extreme QA issues are fine-grained, from what I read. Sadly, like |
110 |
yourself, I like to put on my 'common sense' glasses after an |
111 |
architectural solution is presented, and I've seen mountains of bad |
112 |
ideas; like BP running prudhoe bay (N. Americas largest oil field) in |
113 |
the Arctic. Bad, bad idea, if you are an engineer and hang out with |
114 |
those 'tards' a few days. Collected data in the arctic, microwaved it to |
115 |
a mainframe in Anchorage, ran software, and then microwave controls |
116 |
signals back to the field controllers. Beyond stupid.They were an |
117 |
embarrassment to the entire petroleum industry back in the 70s, when I |
118 |
did some automation (RF to RF) to mainframe work in the arctic. LIke |
119 |
wise the solution to all of the drilling disasters, world wide, is each |
120 |
country needs to provide RT date to a monitoring station, in the |
121 |
government and status things like the condition of the safety and backup |
122 |
safety systems (Real Time) so keep mid manager from making gargantuanly |
123 |
stupid decisions. There is more than this amount of stupidity in how |
124 |
many cluster (cloud companies) think large amounts of critical data will |
125 |
be 'outsourced'. Bean counters scare me the most. |
126 |
Sales-lizards are rarely trusted, unless they listen to me and do |
127 |
exactly what I tell them to do. |
128 |
|
129 |
It seems that there are many many tards in the cluster (cloud) space |
130 |
lacking of common sense. So that (cluster/cloud) industry is going to |
131 |
implode, just like the "dot-com" bubble of the 90s. Not because there is |
132 |
not lots of valid projects and good ideas, but many tards are managing |
133 |
and they lack the common sense to poor piss out of a boot let alone |
134 |
discern valid solutions for specific industries. Like a 'blind hog':: |
135 |
though they will find an acorn or two. A historical CS class or two on |
136 |
what has been tried what works and does not work and why, along with a |
137 |
few (real) hardware architecture classes) and there would not be so many |
138 |
ridiculous (doomed to fail before getting stared) cluster (cloud) |
139 |
companies out there. Developing unknown but old ideas in java, is still |
140 |
going to fail. Many are the BP of the cloud:: a disaster just waiting to |
141 |
fail.... ymmv. Many folks in the Petroleum industry warned Alaskan |
142 |
government officials that BP was incompetent, back in the 70s. |
143 |
They still are mostly becase the executives would not not how to |
144 |
calculate the weight of drill stem column of fluid and match it up with |
145 |
the expected subsurface pressures to be encountered. It's a simple |
146 |
'material balance equation' you could teach a HS physics class. |
147 |
|
148 |
Likewise there is a rich history (graveyard) of distributed processing |
149 |
and that body of knowledge is being ignore, mostly because it is |
150 |
getting in the way of vendor hyperbole...... |
151 |
|
152 |
|
153 |
Douglas did manage to pull his own bacon from the fire, in the end of |
154 |
his article, but it wreaks of vendor hyperbole, imho. |
155 |
|
156 |
|
157 |
thanks for the comments, |
158 |
James |