How much time does x86 Engineering has? Just curious to see what your thoughts are. I think MH will make sure they are gone by the end of 2018.
13 replies (most recent on top)
For example, they realized in short order, that sourcing hardware from outside vendors ended up costing way more in money and time and support contracts and support delays than the initial cost savings. We are now filling up our data centers with our own hardware, with our own in-house support and development engineers to create, maintain, and support that hardware.
I'm hoping you mean they figured that out at least 8 years ago, because that's when Amazon, Facebook -- yeah, and even Sun! -- figured it out. Maybe earlier.
This is what I don't understand: When Oracle Cloud first got started, despite the SPARC/Solaris team trying to get a word in edgewise, a SPARC offering was never considered. When it finally did get considered, for some reason it got tied to Sonoma.
I mean -- what the heck?
The whole idea behind cloud is you go for economy of scale in the back, with the ability to dynamically slice and dice workloads for your front-end offerings. This is what big SPARC iron and Solaris were built for. why would you have to wait for smaller systems?
By the time Sonoma was released, it was way too late. Just as it's always been, the signalling was already done and clear: Oracle didn't care enough about SPARC to even have it in the cloud. Which is about what every Oracle corporate message (other than dedicated SPARC/Solaris launches) screamed, starting way before the acquisition.
@P3XPApv-oyl - there is no one size fits all in Cloud. Look at what Facebook does. It's not like they are only using white boxes. Some of their stuff is outright wild. Way more customized than anything either ORCL or HP have ever done.
They also just sold a bunch of SPARC systems to Bloomberg.
They just Sold a bunch of x86 to AT&T
I expect there will be some small impact on x86 with the layoffs today.
does it matter to you now?
@P3XPApv-oyl SPARC has at least some advantages (robust 16S infrastructure, SQL accelerator, good memory bandwidth). But x86 are just regular 2S x86 systems tied together with added networking (not supported in silicon). You can ask HP and you will get much better x86 server for much better price.
@P3XPApv-api, you are dillusional. x86 group is adding zero value, will be gone in 1-2 years.
Try applying for those openings. These are all stale. The skeptical side of thinks these are left up to create a false sense of security, but it's more likely that there has not been a priority on cleaning them up.
There are openings for hardware in SCA x86. If u go to self service and click on recruiting and search. Unbelievable.
"The key question to ask is what value do they have to Oracle's cloud strategy?"
That statement is absolutely positively 100% correct, and it is the basis for everything and anything that Oracle is doing right now, whether it is hardware, software, people, whatever.
In order to be a successful cloud player you have to have data centers - lots of them, all over the world, and having data centers means that you need lots of hardware, software, and people to fill them up and make them run.
The management chain in charge of these global data centers is not what I would call your typical do-nothing, don't have a clue, just along for the ride Oracle managers. For example, they realized in short order, that sourcing hardware from outside vendors ended up costing way more in money and time and support contracts and support delays than the initial cost savings. We are now filling up our data centers with our own hardware, with our own in-house support and development engineers to create, maintain, and support that hardware.
Oh, and for the record, I have been with Oracle for over 10 years in various roles, and I now work in one of our cloud data centers, so I'm not just talking out of my a**.
The key question to ask is what value do they have to Oracle's cloud strategy? There is no general purpose server business anymore. They will need to sustain engineered systems. How much of a hardware team do they need for their cloud business? I suspect a small fraction of what they have now. I would think at most they would need a small team that would with management of external providers.
Days are numbered. It will eventually go maintenance mode and they will sell it as long as customers have an appetite for engineered systems. Same goes for SPARC. They will still sell it as long as customers are willing to pay for it. Just don't expect the technology to progress other than getting faster chips, more memory, and higher capacity disk drives.