1 |
On 08/04/2016 05:09 AM, J. Roeleveld wrote: |
2 |
> On Tuesday, August 02, 2016 12:16:32 AM james wrote: |
3 |
>> On 08/01/2016 11:49 AM, J. Roeleveld wrote: |
4 |
>>> On Monday, August 01, 2016 08:43:49 AM james wrote: |
5 |
> |
6 |
> <snipped> |
7 |
> |
8 |
>>>> Way back, when the earth was cooling and we all had dinosaurs for pets, |
9 |
>>>> some of us hacked on AT&T "3B2" unix systems. They were know for their |
10 |
>>>> 'roll back and recovery', triplicated (or more) transaction processes |
11 |
>>>> and 'voters' system to ferret out if a transaction was complete and |
12 |
>>>> correct. There was no ACID, the current 'gold standard' if you believe |
13 |
>>>> what Douglas and other write about concerning databases. |
14 |
<snip> |
15 |
>> Comparing results of codes run on 3 different processors or separate |
16 |
>> machines for agreement withing tolerances, is quite different. The very |
17 |
>> essence of using voting where there a result less that 1.0 (that is |
18 |
>> n-1/n or n-x/n was requisite on identical (replicated) processes all |
19 |
>> returning the same result ( expecting either a 0 or 1) returned. Results |
20 |
>> being logical or within rounding error of acceptance. Surely we need not |
21 |
>> split hairs. I was merely pointing out that the basis telecom systems |
22 |
>> formed the early and of widespread transaction processing industries and |
23 |
>> is the grand daddy of the ACID model/norms/constructs of modern |
24 |
>> transaction processing. |
25 |
> |
26 |
> Hmm... I am having difficulty following how ACID and ensuring results are |
27 |
> correct by double or triple checking are related. |
28 |
|
29 |
Atomicity; Consistency; Isolation, Durability == ACID (so we are all on |
30 |
the same page). |
31 |
|
32 |
Not my thesis. My thesis, inspired by these threads, is that all of |
33 |
these (4) properties of ACID, originated in the telephone networks, as |
34 |
separate issues. When telephonic switching moved from electro-mechanical |
35 |
systems to computers, each of these properties where develop by the |
36 |
telephonic software and equipment providers. Banks followed the |
37 |
switching systems and these (4) ACID properties were realized to be |
38 |
universally useful and instituted and rebranded as 'transactions' |
39 |
|
40 |
Database systems, developed by IBM and other quickly realized the value |
41 |
of ACID properties in all sorts of forms of data movement and |
42 |
modification (ie the transaction). Database developers and vendors |
43 |
did not invent ACID properties. Indeed and in fact those properties were |
44 |
first used collectively in the legacy telephonic systems, best |
45 |
desribed by SS(7). Earlier version are a case study in redundancy and |
46 |
reliability of those early telecom systems. Granted latency was a big |
47 |
problem, that moving from electric circuits to digital circuits was |
48 |
fixed; yet still there was the five-nines of quality (99.999%) wonderful. |
49 |
|
50 |
|
51 |
>> For massively parallel needs, |
52 |
>> distributed processing rules, but it is not trivial |
53 |
> |
54 |
> Agreed. |
55 |
|
56 |
<snip> |
57 |
|
58 |
> |
59 |
>> Another point, there are single big GPUs that can be run as thousands of |
60 |
>> different processors on either FPGA or GPU, granted using SIMD/MIMD |
61 |
>> style processors and thing like 'systolic algorithms' but that sort of |
62 |
>> this is out of scope here. (Vulcan might change that, in an open source |
63 |
>> kind of way, maybe). Furthermore, GPU resources combined with DDR-5 can |
64 |
>> blur the line and may actually be more cost effective for many forms of |
65 |
>> transaction processing, but clusters, in their current forms are very |
66 |
>> much general purpose machines. |
67 |
> |
68 |
> I don't really agree here. For most software, having a really fast CPU helps. |
69 |
> Having a lot of mediocre CPUs means the vast majority isn't doing anything |
70 |
> useful. |
71 |
> Software running on clusters needs to be written with massive parallel |
72 |
> processing in mind. Most developers don't understand this part. |
73 |
|
74 |
Where did you get the idea that folks builing clusters, are not as |
75 |
interested in using the fastest processors possible; dude, that's just |
76 |
failed (non-sequitur)logic. |
77 |
|
78 |
Well this premise of yours is a corollary to my thesis; and the early |
79 |
telecom systems developers were historically 'bad ass' and highly |
80 |
intelligent. It has taken the software development world decades to |
81 |
catch up to key systems attributes of hardware design (redundancy and |
82 |
roll-back and recovery). Now that things are digital, you can run codes |
83 |
on a variety of different hardware to abstract the properties of ACID |
84 |
and supercede ACID, with yet more properties of robust hardware design. |
85 |
(Sadly, even most EE professors are severly lacking in this knowledge). |
86 |
Modern EE experts have most of their magic attributed to European |
87 |
Mathmeticians, but that's another issue, too complex for the average |
88 |
java* coder. Curiously, you can read all about, Hilbert, should you need |
89 |
to scratch that itch.... |
90 |
|
91 |
|
92 |
|
93 |
>> My point:: Douglas is dead wrong about ACID being dominated by Databases, |
94 |
>> for technical reasons, particularly for advanced teams of experts. |
95 |
> |
96 |
> Wikipedia actually disagrees with you: |
97 |
> https://en.wikipedia.org/wiki/ACID |
98 |
> "In computer science, ACID (Atomicity, Consistency, Isolation, Durability) is |
99 |
> a set of properties of database transactions." |
100 |
|
101 |
Exactly. Database vendors got the ideas and components (literals and |
102 |
abstractions |
103 |
from the telephonics industries to get a leg up on moving electronic |
104 |
switching (which already had those key components now referred to as |
105 |
ACID) in hardware. When those electro-mechanical systems move to digital |
106 |
circuits, Bell labs ensure those properties where a closely hend secret |
107 |
wrapped up in the 'unix OS' They did promote ACID in their software and |
108 |
the banks were the other customers were likewise saying YES YES YES, we |
109 |
want telcom ACID level of performance in our (developing) computer |
110 |
software too. But the migration to digital let the 'cat out of the bag' |
111 |
on the wonders of ACID (long before Timothy Leary, just so the |
112 |
Californians among us can keep up!). |
113 |
|
114 |
|
115 |
> In other words, it's related to databases |
116 |
|
117 |
They (vendors) copied it from telecom, and wildly promoted it, very |
118 |
successfully. Combine this with the fact that most US EE programs are |
119 |
abysmally weak (always have been), so now we indeed and in fact have |
120 |
this severe lapse in robust and fault tolerant systems. |
121 |
|
122 |
WHY? Nothing (industrial or commercial) had the "Five-nines" of |
123 |
reliability, but those electro-mechanical telephonic systems. |
124 |
*nothing* Everybody wanted it; hence those (4) components were harvested |
125 |
from telephonics and used as a model for all transactions. |
126 |
|
127 |
Take "atomicity" for example. It has it's roots in "call setup". |
128 |
Dialogic is a pc board vendor (from decades ago) that followed those |
129 |
early systems. Here is a document (from the 70s/80s/?) were they |
130 |
have "40 Atomic Functions" that they use in software to control the |
131 |
hardware for 'call setup and management'. Sure many more documents |
132 |
exist, but they may not be publically available in electronic forms. |
133 |
All of this occurred before those folks that write for Wikipedia were |
134 |
ever born, so they could not possible be aware of these issues and |
135 |
historical precedence. |
136 |
|
137 |
[1] |
138 |
https://www.dialogic.com/webhelp/MSP1010/10.4.0/WebHelp/ppl_dg/l3p_cic.htm |
139 |
|
140 |
|
141 |
One can research each of those four properties and discover how telecom |
142 |
integrated them into the phone system of North America (Europe almost |
143 |
evolved simultaneously). Bell Labs is " the data of ACID"; and it was a |
144 |
tightly held secret as long as possible, to delay the expansion of usage |
145 |
and eventual break up of that legacy monopoly. |
146 |
|
147 |
There are many things in the (legacy) communications world that have not |
148 |
accurately made it's way to digital in a form freely available on the |
149 |
internet. (like signal intercept). Think of all of those hidden |
150 |
antennae arrays in the UK when microwave telecom was all the rage. |
151 |
MCI was a key player on exploiting microwave (another tenant of EE). |
152 |
|
153 |
|
154 |
>> Surely most MBA, HR and Finance types of |
155 |
>> idiots running these new startups would know know a coder from an |
156 |
>> architect, and that is very sad, because a good consultant could have |
157 |
>> probably designed several robust systems in a week or two. Grant few |
158 |
>> consultants has that sort of unbiased integrity, because we all have |
159 |
>> bills to pay and much is getting outsourced... Integrity has always been |
160 |
>> the rarest of qualities, particularly with humanoids...... |
161 |
> |
162 |
> The software Uber uses for their business had to be developed in-house as |
163 |
> there, at least at the time, was nothing available they could use ready-made. |
164 |
> This usually means, they start with something simple they can get running |
165 |
> quickly. If they want to fully design the whole system first, they would never |
166 |
> get anything done. |
167 |
> |
168 |
> Where these projects usually go wrong is that they wait too long with a good |
169 |
> robust design, leading to a near impossibility of actually fixing all the, in |
170 |
> hindsight obvious, design mistakes. |
171 |
> (NOTE: In hindsight, as most of the actual requirements would not be clear on |
172 |
> day 1) |
173 |
|
174 |
I could not agree with you more. |
175 |
|
176 |
The more processors, readily available to codes that know how to use |
177 |
them, in parallel the faster and better and more reliable the systems |
178 |
developed (including the software) will be. Some are working on extremly |
179 |
low latency systems where FPGAs are embedded in general purpose |
180 |
processors (Intel is leading on this). The DoD has been using these |
181 |
systems for decades. Clusters are superior to single (or multicore) |
182 |
systems if these kids knew anything about redundancy and fault |
183 |
tolerance; both which originate in hardware and the telecom industries |
184 |
perfected to the 99.999% robustness level (while IBM drulled on their |
185 |
punch-cards. I know, I was there...... |
186 |
|
187 |
And in my opinion,that was the most important of the collective of |
188 |
reasons why AT&T, it's 10,000+ lawyers and assholes in our government |
189 |
fought so hard to keep early unix expansion out of the hands of the |
190 |
masses. At one point it was easier to get a top-secret clearance than it |
191 |
was to code on those early telecom systems. |
192 |
|
193 |
|
194 |
>>>> and the switch was was configured, then the code would |
195 |
>>>> essentially 'vote' and majority ruled. This is what led to phone calls |
196 |
>>>> (switched phone calls) having variable delays, often in the order of |
197 |
>>>> seconds, mis-connections and other problems we all encountered during |
198 |
>>>> periods of excessive demand. |
199 |
>>> |
200 |
>>> Not sure if that was the cause in the past, but these days it can also |
201 |
>>> still take a few seconds before the other end rings. This is due to the |
202 |
>>> phone-system (all PBXs in the path) needing to setup the routing between |
203 |
>>> both end-points prior to the ring-tone actually starting. |
204 |
>>> When the system is busy, these lookups will take time and can even |
205 |
>>> time-out. (Try wishing everyone you know a happy new year using a wired |
206 |
>>> phone and you'll see what I mean. Mobile phones have a seperate problem |
207 |
>>> at that time) |
208 |
>> I did not intend to argue about the minutia of how a particular Baby |
209 |
>> Bell implemented their SS7 switching systems on unix systems. My point |
210 |
>> was the 'transaction processing' grew out the early telephone network, |
211 |
>> the way I remember it:: ymmv. Banks did dual entry accounting by hand |
212 |
>> and had clerks manually load data sets, then double entry accounting |
213 |
>> became automated and ACID style transaction processing added later. So |
214 |
>> what sql folks refer to as ACID properties, comes from the North |
215 |
>> American switching heritage and eventually the worlds telecom networks, |
216 |
>> eons ago. |
217 |
> |
218 |
> There is a similarity, but where ACID is a way of guaranteeing data integrity, |
219 |
> a phone-switch does not need this. It simply needs to do the routing |
220 |
> correctly. |
221 |
|
222 |
Have you every talked to an old military officer that worked in |
223 |
Intelligence? Like the spy plan incidence over Afganistan, circa 1960 |
224 |
[2]? https://en.wikipedia.org/wiki/1960_U-2_incidentf |
225 |
|
226 |
Data integrity almost caused WW2. |
227 |
|
228 |
WRONG. The fives-nines was so coveted by everyone else that there was a |
229 |
feeding frezy on just how these folks at bell labs pulled it off. Early |
230 |
(1950-1970s) computational systems were abysmal to own or operate and |
231 |
yet the sorry ass phone company had 99.999% perfection (thanks to bell |
232 |
labs)? They provided the T1 and T3 lines in/out of the pentagon. |
233 |
Jealousy was outrageous. Database vendors where struggling with |
234 |
assembler and 'board changeouts' as Rich alluded to. |
235 |
|
236 |
<snip> |
237 |
|
238 |
>>> ACID is about data integrity. The "best 2 out of 3" voting was, in my |
239 |
>>> opinion, a work-around for unreliable hardware. |
240 |
|
241 |
Correct. voting was used as the precursor technology to distributed |
242 |
systems (today it's the cluster), It added to the reliablity and |
243 |
robustness. It provided consistency. It demonstrated that the entire |
244 |
string of what was need for ss7, including call setup, could be |
245 |
replicated and run on a cluster (oops another hardware set).... |
246 |
|
247 |
|
248 |
>> Absolute true. But the fact that a High Reliability in computer |
249 |
>> processing (including the billing) could be replicated performed |
250 |
>> elsewhere and then 'recombined', proves that the need of any ACID |
251 |
>> function can be split up and ran on clusters and achieve ACID standards |
252 |
>> or even better. So my point, is that the cluster, if used wisely, |
253 |
>> will beat the 'dog shit' out of any Oracle fancy-pants database |
254 |
>> maneuvers. Evidence:: Snoracle is now snapping up billion dollar |
255 |
>> companies in the cluster space, cause their days of extortion are |
256 |
>> winding down rather rapidly, imho. |
257 |
|
258 |
> I disagree here. For some workloads, clusters are really great. But SQL |
259 |
> databases will remain. |
260 |
|
261 |
As a subset of distributed processing. Oracle (the champion of |
262 |
databases) is going to atrophy and slip into irrelevance, once kids |
263 |
learn how to supersede ACID with judicious cluster hardware and codes on |
264 |
top of heterogeneous clusters..... Granted any corp with billions and |
265 |
billions and deep (illegal?) relationships with government officals will |
266 |
eventually prosper again.... |
267 |
|
268 |
Once again, EE will light the forward path. |
269 |
|
270 |
>> Also, just because the kids are writing the codes, have not figured all |
271 |
>> of this out, does not mean that SQL and any abstraction is better that |
272 |
>> parallel processing. No way in hell. Cheaper and quicker to set up, |
273 |
>> surely true, but never superior to a well design properly coded |
274 |
>> distributed solution. That's my point. |
275 |
> |
276 |
> Workloads where you can split the whole processing into small chunks where the |
277 |
> same steps can be performed over a random sized chunk and merging at a later |
278 |
> stage will lead to correct results. Then yes. |
279 |
|
280 |
True, but it's not quite as restrictive as you think. Large system, |
281 |
with even just a small bit of parallism integrated into the overall |
282 |
architecture, benefit. Howmuch depends on the designers. We do need |
283 |
more EE coders leading on cluster designs, but the Universities (world |
284 |
wide) have let everyone down. |
285 |
|
286 |
|
287 |
|
288 |
> However, I deal with processes and reports where the amount of possible chunks |
289 |
> is definitely limited and any theoretical benefit of splitting it over multiple |
290 |
> nodes will be lost when having to build a very fancy and complex algorithm to |
291 |
> merge all the seperate results back together. |
292 |
|
293 |
NoSQL is an abysmal failure. SQL need to be a small subset of robust |
294 |
parallel systems design and implementation. The latest venue is |
295 |
'unikernels'. |
296 |
|
297 |
Cluster will dominate because deep pockets can have the latest and |
298 |
fastest and cheapest hardware, in massive quantities before the |
299 |
commoners even learn how it works. Arm64V8 is a prime example and |
300 |
current example. It's heat loading per unit of processing, blows away |
301 |
Cisc based systems. FPGA can implement any processor or memory structure |
302 |
and can it in microseconds. But these are areas where attornies via the |
303 |
patent system, abuse light-weight competition. |
304 |
|
305 |
|
306 |
> This algorithm then also needs to be extensively tested analysed and |
307 |
> understood by future developers. The additional cost involved will be |
308 |
> prohibitive. |
309 |
|
310 |
Don't we need more jobs? Are you kidding me? That's way large |
311 |
corporations are so vehemently aggressive in these spaces. We have all |
312 |
kinds of 'stem graduates' here in the US that cannot get a stem job. |
313 |
(hence trumps appeal to the middle class:: tarrifs and promote |
314 |
competition at home). |
315 |
|
316 |
|
317 |
> I disagree, UBER is still using a relational database as the storage layer |
318 |
> with something custom put over it to make it simpler for the developers. |
319 |
> Any abstraction layer will have a negative performance impact. |
320 |
|
321 |
Wanna bet that UBER and like minded companies change again and again and |
322 |
again, until they start study of what mathematicians and EE have been |
323 |
doing for a very long time. |
324 |
|
325 |
>>> It is based on a clever idea, but when |
326 |
>>> 2 computers having the same data and logic come up with 2 different |
327 |
>>> answers, I wouldn't trust either of them. |
328 |
|
329 |
This is rare occurance in digital systems. However, when you look at |
330 |
other forms of computational mathematics, tolerances have to be used |
331 |
to get consistency (oops another property of acid showing up in legacy |
332 |
literature). |
333 |
|
334 |
I could not care less about UBER's problems, unless they send some funds |
335 |
my way. BUT, I am willing to share knowledge, so they 'wise up' because |
336 |
fundamentally, I love disruption in the status quo. |
337 |
|
338 |
>> |
339 |
>> Yep, That the QA of Transactions is rejected and must be resubmitted, |
340 |
>> modified or any number of remedies, is quite common in many forms of |
341 |
>> software. Voting does not correct errors, except maybe a fractional |
342 |
>> rounding up to 1(pass) or down to zero (failure). It does help to |
343 |
>> achieve the ACI of ACID |
344 |
> |
345 |
> It's one way of doing it. But it can also cause extra delays due to having to |
346 |
> wait for seperate nodes to finish and then to check if they all agree. |
347 |
|
348 |
Once clusters are prototyped on Cisc systems, Those codes will be |
349 |
rapidly moving to DSPs, GPUs and FPGA and DDR5+. Those with deep pockets |
350 |
will 'smoke' the competition and idiots like Verizon |
351 |
will be trying to make more stupid acquisitions. Folks do know that |
352 |
Verizon sold off billions in data centers, close to fiber highway |
353 |
to by Yahoo, right? (It "pays out" because they are actually dumping |
354 |
hundreds of thousands of legacy employees (trump voters); that's what |
355 |
that transaction is all about. They are still doom to fail, because the |
356 |
software idiots advising Verizon, have no clue about the fundamentals |
357 |
and mathematics of Communications. (very sad state of affair for Verizon). |
358 |
|
359 |
|
360 |
>> Since billions and billions of these (complex) transactions are |
361 |
>> occurring, it is usually just repeated. If it keeps failing then |
362 |
>> engineers/coders take a deeper look. Rare statistical anomalies are |
363 |
>> auto-scrutinized (that would be replications and voting) and the pushed |
364 |
>> to a logical zero or logical one. |
365 |
> |
366 |
> The complexity comes from having to mould the algorithm into that structure. |
367 |
> And additional complexity also makes it more fault-likely. |
368 |
|
369 |
Only during development and beta tests. After a while it will become |
370 |
'rock solid' and pushed down into the lowest levels of hardware, so it |
371 |
is hidden from the average coder. Here is a billionare, who is quite |
372 |
stealthy, that has done this exact thing most recently. |
373 |
|
374 |
[3] https://www.deshawresearch.com/ |
375 |
[4] |
376 |
https://www.quora.com/unanswered/Computer-Architecture-How-its-like-working-for-DESHAW-RESEARCH-as-an-ASIC-designer-architect |
377 |
|
378 |
<snip> |
379 |
|
380 |
> A lot can be described using 'modern' designs. However, the fact remains that |
381 |
> ACID was worked out for databases and not for phone systems. Any sane system |
382 |
> will have some form of consistency checks, but the extent where this is done |
383 |
> for a data storage layer, like a database, will be different to the extent |
384 |
> where this is done for a switching layer, like a router or phone switch. |
385 |
|
386 |
Please reread my previous posts. You, or anyone can do the individual |
387 |
(and robust) research on the ACID components and the history of telecom. |
388 |
|
389 |
Wikipedia and many other sites have failed you here; sorry. |
390 |
|
391 |
<snip> |
392 |
|
393 |
> Those incompetencies are usually in the domain of finances and services |
394 |
> provided. The basic service of a telecoms company is pretty simple: "Pass |
395 |
> data/voice between A and B". |
396 |
> There are plenty of proven systems available that can do this. The mistakes |
397 |
> are usually of the kind: The system that we bought does not handle the load |
398 |
> the salesperson promised. |
399 |
|
400 |
ON the surface, you are absolutely correct. Mass education is severly |
401 |
thrwated by the entire patent system, grotesque lawyers and legal |
402 |
semantics and the 'bought and sold politicians' from around the globe. |
403 |
(the same folks that brought us globalism). So folks are merely |
404 |
"uneducated" in these matters. Yes these globalists continue to consipre |
405 |
against commoners, around the globe. Education and sharing of hardware |
406 |
and software and mathematics and physics will set the captives free |
407 |
(eventually). This is the essence of WW3 imho. |
408 |
|
409 |
The fact that the masses and even most coders are blissfully unaware of |
410 |
where ACID came from, is a testament to the failure of globalism that |
411 |
provides the protection to the billionaire class of manipulators, imho. |
412 |
|
413 |
> |
414 |
>>> With a small number, it might actually still scale, but when you pass a |
415 |
>>> magic number (no clue what this would be), the counting time starts to |
416 |
>>> exceed any time you might have gained by adding more voters. |
417 |
>> |
418 |
>> Nope the larger the number, the more expensive. The number of voters |
419 |
>> rarely goes above 5, but it could for some sorts of physics problems |
420 |
>> (think quantum mechanics and logic not bound to [0 1] whole numbers. |
421 |
>> Often logic circuits (constructs for programmers, have "dont care" |
422 |
>> states that can be handled in a variety of ways (filters, transforms, |
423 |
>> counters etc etc). |
424 |
> |
425 |
> "don't care" values should always be ignored. Never actually used. (Except for |
426 |
> randomizer functionality) |
427 |
|
428 |
Dude, you need to find some Rf/analog folks and learn about what's going |
429 |
on around "noise" in systems. Once thought to be useless, or a |
430 |
hindrance, it is a fertile ground for innovation, again that the masses |
431 |
are blissfully unaware of. Much is termed "classified" just so you know. |
432 |
|
433 |
> |
434 |
>>> Also, this, to me, seems to counteract the whole reason for using |
435 |
>>> clusters: |
436 |
>>> Have different nodes handle a different part of the problem. |
437 |
>> |
438 |
>> That also occurs. But my point is properly design code for the cluster |
439 |
>> can replace ACID functions, offered by Oracle and other over priced |
440 |
>> solutions, on standard cluster hardware. |
441 |
> |
442 |
> All commonly used relational databases have ACID functionality as long as they |
443 |
> support transactions. There is no need to only choose a commercial version for |
444 |
> that. |
445 |
|
446 |
Like the Chinese, they are brilliant copy cats:: nothing wrong with that |
447 |
(see my take on 100% absolution of all patents, globally. |
448 |
|
449 |
> |
450 |
>> The problem with todays |
451 |
>> clusters is the vendors that employ the kid-coders, are making things |
452 |
>> far more complicated that necessary, so the average linux hacker just |
453 |
>> outsources via the cloud. DUMB, insecure and not a wise choice for many |
454 |
>> industries. |
455 |
> |
456 |
> Moving your entire business into the cloud often is. |
457 |
I could not agree more. HYBRID systems, where the chief |
458 |
architect/designer works exclusively for the custer, is where the future |
459 |
will shake out. All of this idiocy on the masses on the web:: who cares |
460 |
where it is processed. The closer to the node-idiot-user-consumer, the |
461 |
better, mathematically. |
462 |
|
463 |
> |
464 |
>> And sooner or later folks are going to get wise can build |
465 |
>> their own clusters that just solve the problems they have. Surely hybrid |
466 |
>> clusters will domiant where the owner of the codes does outsource peak |
467 |
>> loads and mundance collects of ordinary (non-critical) data. |
468 |
> |
469 |
> Eg. hybrid solutions... |
470 |
|
471 |
Yes yes and HELL YES! In fact gentoo stands out for the quintessential |
472 |
'unikernel' for distributed processing! |
473 |
|
474 |
|
475 |
>> Vendors know this and have started another 'smoke and mirrors' campaign called |
476 |
>> (brace yourself) 'Unikernels'..... |
477 |
> |
478 |
> "unikernels" is something a small group came up with... I see no practical |
479 |
> benefit for that approach. |
480 |
|
481 |
A minimize gentoo system and an optimize and severly stripped linux |
482 |
kernel is pretty much a unikernel. Docker, the leader in |
483 |
commercialization of containers, knows this and has subsummed Alpine |
484 |
linux. Patients my friend, it will become very clear over time, but not |
485 |
exactly the way the current vendors are portraying unikernels. |
486 |
|
487 |
> |
488 |
>> Problem with that approach is they |
489 |
>> should just be using minized (focused) gentoo on striped and optimize |
490 |
>> linux kernels; but that is another lost art from the linux collection |
491 |
> |
492 |
> I see "unikernels" as basically, running the applications directly on top of a |
493 |
> hypervisor. I fail to see how this makes more sense than starting an |
494 |
> application directly on top of an OS. The whole reason we have an OS is to |
495 |
> avoid having to reinvent the wheel (networking, storage, memory handling,....) |
496 |
> for every single program. |
497 |
|
498 |
(see above response). For the last few years, I have run into an |
499 |
astounding number of brilliant folks that have mastered and use gentoo |
500 |
on a daily basis. The more I learn about clusters, the more I realize |
501 |
why this massive of gentoo folks are so silent on these matters. |
502 |
Strategic business plans, brah. Gentoo is the worlds best kept secret. |
503 |
|
504 |
> |
505 |
>>> Clusters of multiple compute-nodes is a quick and "simple" way of |
506 |
>>> increasing the amount of computational cores to throw at problems that |
507 |
>>> can be broken down in a lot of individual steps with minimal |
508 |
>>> inter-dependencies. |
509 |
>> |
510 |
>> And surpass the ACID features of either postgresql or Oracle, and spend |
511 |
>> less money (maybe not with you and postgresql on their team)! |
512 |
> |
513 |
> Large clusters are useful when doing Hadoop ("big data") style things (I |
514 |
> mostly work with financial systems and the corresponding data). |
515 |
> Storing the entire datawarehouse inside a cluster doesn't work with all the |
516 |
> additional requirements. Reports still need to be displayed quickly and a |
517 |
> decently configured database is usually more beneficial. Where systems like |
518 |
> Exadata really help here is by integrating the underlying storage (SAN) with |
519 |
> the actual database servers and doing most of the processing in-memory. |
520 |
> Eg. it works like a dedicated and custom build cluster environment specifically |
521 |
> for a relational database. |
522 |
|
523 |
There is a revolution in hardware memory technologies. In a few more |
524 |
years massive ram will be an integral part of of the computational |
525 |
hardware (think DDR5 and GPUs currently. Most massive systems can be |
526 |
split up into small systems too. Databse vendors have little incentive |
527 |
to do this for customers. The art of the design and implementation of |
528 |
'transaction processing' need to return to hardware concepts during this |
529 |
transition. |
530 |
|
531 |
> |
532 |
> |
533 |
>>> I say "simple" because I think designing a 1,000 core chip is more |
534 |
>>> difficult than building a 1,000-node cluster using single-core, single |
535 |
>>> cpu boxes. |
536 |
>> Today, you are correct. Tomorrow you will be wrong. |
537 |
> |
538 |
> In that case, clusters will be obsolete tomorrow. |
539 |
|
540 |
No, the chips and the cluster will be one in the same. Real time |
541 |
sequence stepping in problem->solution domains for things like flight |
542 |
simulation and subsurface fluid management are still grand challenges |
543 |
that are a ways off. The average database solution, even for large |
544 |
commercial/global operations, is going to migrate to clusters. Clusters |
545 |
and storage will continue to migrate to silicon. The biggest problem is |
546 |
the patent system and artificial constructs more commonly known in the |
547 |
business world as "cost barrier to entry" economics. These mostly result |
548 |
from the way the local/state/federal/global laws are implemented and |
549 |
enforced. |
550 |
|
551 |
> |
552 |
>> [1]. Besides once |
553 |
>> that chip or VHDL code or whatever is designed, it can be replicated and |
554 |
>> resused endlessly. Think ASIC designers, folks to take a fpga project to |
555 |
>> completing, An EE can codes on large arrays of DSPs, or a GPU |
556 |
>> (think Khronos group) using Vulcan. |
557 |
>> |
558 |
>>> I would still consider the cluster to be a single "machine". |
559 |
>> |
560 |
>> Thats the goal. |
561 |
> |
562 |
> That, in my opinion, that goal has already been achieved. Unless you want ALL |
563 |
> machines to be part of the same cluster and all machines being able to push |
564 |
> work to the entire cluster... |
565 |
> In that case, good luck in achieving this as you then also need to handle |
566 |
> "randomly dissapearing nodes" |
567 |
|
568 |
I think Brexit and Trump will replace globalism with localism and |
569 |
tariffs. Goverments will fight over the spoils of tariffs to finance |
570 |
their glutony, and locals will figure out how to build and operate |
571 |
everything, locally. So you are correct. I actually am promoting hybrids |
572 |
clusters, so the commoners can ;'suck the brain-marrow' out of |
573 |
walstreet, politicans and the globalists. Once groups of locals learn to |
574 |
be self sufficient, think of them and digital omish, the only function |
575 |
governemnts and globallist provide is national security. Folks that like |
576 |
work can join up and kills folks from other like minded collectives. |
577 |
Most will be extraordinarily happy to provide 100% of what they need, |
578 |
locally. There will be some exchange of material and those less |
579 |
innovative will lag a bit, but that is what globalist should concentrate |
580 |
on:: how to teach those less fortunate how to become self sufficient, |
581 |
locally. |
582 |
|
583 |
|
584 |
> And 90+% of developers still don't understand how to properly code for multi- |
585 |
> threading. Just look at how most applications work on your desktop. They all |
586 |
> tend to max out a single core and the other x-1 cores tend to idle... |
587 |
|
588 |
Wonder why Bill Gates (in his tax-dogging world charities) is not |
589 |
teaching this stuff? Rupert Murdock? Rich Arabs? Chineese? |
590 |
|
591 |
The elites of the world are 'selfish bastards' and use the good work |
592 |
that come from their ranks to further screw up localism (self |
593 |
sufficiency on a local basis). Sooner or later these globalist will have |
594 |
to answer to the masses of local citizens, wherever they are hiding. We |
595 |
have seen the purging of the Republican party. The Democratic Elites are |
596 |
currently undergoing a purging. After Brexit, it |
597 |
will rapidly expand in Europe. Saudis are running scared. Pandemic |
598 |
of locals that want to be self sufficient. Folks are tire of listing to |
599 |
some (asshole) expert that does not live down the street from them. |
600 |
Globalism flies in the face of common-sense, and computational |
601 |
competence is not except. There is latency and much deceptions in the |
602 |
work of computations, but that too will fall (eventually). |
603 |
|
604 |
> |
605 |
>> Granted many old decreped codes had to be |
606 |
>> redesigned and coded anew with threads and other modern constructs to |
607 |
>> take advantage of newer processing platforms. |
608 |
> |
609 |
> Intel came with Hyperthreading back in 2005 (or even before). We are now in |
610 |
> 2016 and the majority of code is still single-threaded. |
611 |
> The problem is, the algorithms that are being used need to be converted to |
612 |
> parallel methods. |
613 |
> |
614 |
>> Sure the same is true with |
615 |
>> distributed, but it's far closer than ever. The largest problem with |
616 |
>> cluster, is Vendors with agendas, are making things more complicated |
617 |
>> than necessary and completely ignoring many fundamental issues, like |
618 |
>> kernel stripping and optimizations under the bloated OS they are using. |
619 |
> |
620 |
> I still want a graphical desktop with full multi media support. I still want |
621 |
> to easily plugin a USB device or SD-card and use it immediately,..... |
622 |
> That requirement is incompatible with stripping the OS. |
623 |
|
624 |
Agreed. And I want to build the hardware on my own 3D printer. I am |
625 |
flexible to try out many offerings when 3D printing looses those patents |
626 |
on using metals and semiconductor materials...... |
627 |
This too will come, hopefully sooner than later and without the shedding |
628 |
of blood.... |
629 |
|
630 |
> |
631 |
>>>> I do |
632 |
>>>> not have the old articles handy but, I'm sure that many/most of those |
633 |
>>>> types of inherent processes can be formulated in the algebraic domain, |
634 |
>>>> normalized and used to solve decisions often where other forms of |
635 |
>>>> advanced logic failed (not that I'm taking a cheap shot at modern |
636 |
>>>> programming languages) (wink wink nudge nudge); or at least that's how |
637 |
>>>> we did it.... as young whipper_snappers bask in the day... |
638 |
>>> |
639 |
>>> If you know what you are doing, the language is just a tool. Sometimes a |
640 |
>>> hammer is sufficient, other times one might need to use a screwdriver. |
641 |
>>> |
642 |
>>>> --an_old_farts_logic |
643 |
>>> |
644 |
>>> Thinking back on how long I've been playing with computers, I wonder how |
645 |
>>> long it will be until I am in the "old fart" category? |
646 |
>> |
647 |
>> Stay young! I run full court hoops all the time with young college |
648 |
>> punks; it's one of my greatest joys in life, run with the young |
649 |
>> stallions, hacking, pushing, shoving, slicing and taunting other |
650 |
>> athletes. Old farts clubs is not something to be proud of, I just like |
651 |
>> to share too much...... |
652 |
> |
653 |
> Hehe.... One is only as old as he/she feels. |
654 |
> |
655 |
> -- |
656 |
> Joost |
657 |
|
658 |
|
659 |
Young kids often show amazing wisdom. The educational processes beat |
660 |
this out of kids. Isolation and localism (aka home schooling) does allow |
661 |
kids to explode on both technical competence and creativity. |
662 |
But this flies in the face of the goals of globalism. When I was young, |
663 |
there was a kid that was brilliant and 100% home schooled by mostly |
664 |
uneducated parents. They lived in the bush of Alaska, hundreds of miles |
665 |
from anyone. Brilliance and innovation are the providence of the youth; |
666 |
just look at all of those young, brilliant minds from post-mid-evil |
667 |
Europe. Mass education just beat those traits right out of all children. |
668 |
Communications and localism will yeild many, many brilliant folks and |
669 |
that is the greatest fear of the globalist, who want to remain in power |
670 |
and have dominion over the masses. It's the classic struggle. The path |
671 |
to a better future is espoused in parallel and distributed and local |
672 |
decision/control, from politics to hardware to software. |
673 |
|
674 |
hth, |
675 |
James |