AI and big data with Oracle.

hi,

What is the component/application for Oracle to do big data analysis?

What is the Oracle product for AI and data scientist research ?

when I look at this:

https://www.oracle.com/artificial-intelligence/platform.html

it said:

Additional libraries and tools include: Jupyter, pandas, scikit-learn, Pillow, OpenCV, and NumPy.
Deep learning frameworks include: TensorFlow, Keras and Caffe.
Elastic AI and Machine Learning Infrastructures include NVIDIA, Flash Storage, and Ethernet.

Open in new window

.

which component is free and no issue with Oracle DB. ?
LVL 1
marrowyungSenior Technical architecture (Data)Asked:
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

x
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

slightwv (䄆 Netminder) Commented:
I know very little about AI and what machine learning products are out there.

From the link you provided, those appear to be services you can access from Oracle's Cloud.
0
marrowyungSenior Technical architecture (Data)Author Commented:
this is what i researched yesterday:

1) Big data applicatant. (NoSQL + HDFS, Hadoop run inside)
2) Oracle Exalytics in RAM machine (analysis tools)
3) Exadata. (Oracle DB) with in database analytics.

this is the WHOLE picture, surely all CHARGE separately, right? you have a price list don't you .
0
slightwv (䄆 Netminder) Commented:
>>you have a price list don't you .
Google Oracle Pricing.  They publish their list prices but no one pays that.  Every environment I've ever been in gets some level of discount.

>>this is the WHOLE picture

If you say so.  I'm not sure what all that has to do with the original question about AI.  Yes, Oracle competes in most markets and yes, most products you must pay for.
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
Big Business Goals? Which KPIs Will Help You

The most successful MSPs rely on metrics – known as key performance indicators (KPIs) – for making informed decisions that help their businesses thrive, rather than just survive. This eBook provides an overview of the most important KPIs used by top MSPs.

johnsoneSenior Oracle DBACommented:
I know I'm going to take a lot of flack for this post, but here goes....

Exadata isn't really an Oracle database.  It is a purpose built hardware platform.  Yes, it does run a form of Oracle database on it, but it is difficult to compare to straight database software, as there is specific hardware that goes with it.

If you are considering it, I highly recommend you do a POC with Exadata and it's competitors (all of them).  They are almost all purpose built hardware solutions with different database technology behind them (mostly proprietary).  The problem with them is that it depends on your workload which one is going to be the best for you and that is where a POC comes in.  I don't know if these are still the big players, but it used to be Teradata, Greenplum, IBM Pure Analytics and Oracle Exadata.  All of them used to do POCs and most people were doing them next to one another.

At one point, I worked for a competitor of Exadata, when Exadata was a brand new product.  From the POC data that I had seen, every competitor would beat out Oracle on almost every benchmark.  The only one that Oracle could beat other products at was highly transactional loads, which most people would argue don't belong in big data anyway.

As far as cost...  With big data analytics, the more hardware the better.  You are better off with a system that is way oversized for the data you have.  Seems like a waste of money to overbuy on system size, but it has to do with the degree of parallelism that can be achieved with more hardware.   The more parallelism you have, the faster your queries run.  You can spread your load across 50 drives and 16 processors, but it will run a lot faster across 200 drives and 64 processor (you get the idea).  Pretty much everything in this world is a full table scan, so throwing hardware at it makes it better.  One problem with these systems is they are not easily upgradable.  I know the companies were working on it, but they all used to be forklift upgrades (take out the old system and replace with a new bigger one, nothing in place).

Pretty off topic, but since is seems the question was going this way anyway...
0
marrowyungSenior Technical architecture (Data)Author Commented:
"  One problem with these systems is they are not easily upgradable."

tks.  just like apple stuff.

"(take out the old system and replace with a new bigger one, nothing in place)."

should be

"Seems like a waste of money to overbuy on system size, but it has to do with the degree of parallelism that can be achieved with more hardware. "

yeah. a lot of money on that.
0
marrowyungSenior Technical architecture (Data)Author Commented:
tks
0
johnsoneSenior Oracle DBACommented:
The reason for no upgrade in place is it is very difficult to do.  Think about it.  You have to bring in another rack of equipment and then attach it to all the other racks.  Then redistribute the data.  Takes a really long time and has a tendency to have issues.

Bringing in a whole new system that was put together and already tested, then disassembled and brought on site is a much safer way of doing things.  Especially if the existing system is older and newer versions of hardware are being used, then you have untested configurations, and trying to figure out what went wrong in one of those is impossible.

I know for sure that they were working on trying to do them, but I don't know if they ever got them fully working.  And, I think they would only consider them if you were running the latest hardware.  You couldn't add a rack to a system that was a few years old, things just wouldn't talk to each other correctly.
0
marrowyungSenior Technical architecture (Data)Author Commented:
"The reason for no upgrade in place is it is very difficult to do.  Think about it.  You have to bring in another rack of equipment and then attach it to all the other racks.  Then redistribute the data.  Takes a really long time and has a tendency to have issues."


but you are referring to oracle to oracle ,right ? should be pretty easy !
0
johnsoneSenior Oracle DBACommented:
Not software upgrade.  Hardware upgrade.  Adding another rack of hardware isn't as simple as you may think it is.
0
marrowyungSenior Technical architecture (Data)Author Commented:
" Adding another rack of hardware isn't as simple as you may think it is."

understood.

for MS SQL, quite easy ! procedure quite straight forward ! I do it a lot of time for MS SQL.
0
johnsoneSenior Oracle DBACommented:
It is quite easy for an Oracle database as well.  Not so for Exadata, or any of the other purpose built hardware systems.
0
marrowyungSenior Technical architecture (Data)Author Commented:
"Not so for Exadata, or any of the other purpose built hardware systems."

tks,

but why ?
0
johnsoneSenior Oracle DBACommented:
Because this is not simply a computer connected to some disk drives.  When you add another rack of storage, you are actually adding storage and more computers to the hardware system.  All of the interconnects between the computers need to be set up and tested.  There is a ton of work to be done and from what I know of working for a company that made these, it doesn't go well.  It is a lot easier to build a system from scratch and test and install it than to do an in place upgrade.  Too many things need to be unplugged and too many things need to be plugged in, all in the right ports so everything knows where everything else is.  Even after tons of testing and documentation, the in place upgrades still have problems and run into multi-day processes of debugging.

For instance, the company that I worked for, when you add a rack of storage, what you are actually adding is 8 blades and 128 drives.  These are massively parallel systems, not symmetric multiprocessing (which is the type of parallel processing most people are familiar with), so those 8 new blades need to be able to talk to all the other blades in the system and the 2 controlling computers.
0
marrowyungSenior Technical architecture (Data)Author Commented:
"There is a ton of work to be done and from what I know of working for a company that made these, it doesn't go well."

so this is about HW but not SW, right ? I though you are talking about SW, oracle.

"Too many things need to be unplugged and too many things need to be plugged in, all in the right ports so everything knows where everything else is.  "

this is more like data center migration . cable management is a huge task!
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Big Data

From novice to tech pro — start learning today.