setting up openstack doesn't have to be difficult
I wrote a post about this just yesterday. https://www.linkedin.com/pulse/deploying-openstack-difficult-peter-lees
27 posts • joined 9 Dec 2011
I wrote a post about this just yesterday. https://www.linkedin.com/pulse/deploying-openstack-difficult-peter-lees
Just wanted to re-iterate that, as an open source project, Ceph is not exclusively "owned" by Red Hat (just as it doesn't exclusive "own" Linux).
For example, SUSE Enterprise Storage is based on Ceph. Others also package it up with OpenStack distributions, etc etc.
> As far as the lost sales job, I would be fine if they all were borded on
> Ark B along with the hairdressers and hand sanatizers of Golgafrincham.
...it was telephone sanitisers, and in the end Golgafrincham's population was wiped out by a disease spread through a dirty telephone handset... so be careful what you wish for.
> They are paid to lie, they serve no extisential value other than caring for their families.
> Unfortunately, the disappearing sales jobs will be largely replaced by marketing jobs,
> who also lie, just to a broader audience (because it's all about scalability).
this flawed characterisation is grossly offensive: I'm sorry if you've had some bad experiences, but the vast majority of marketing & sales folk (which includes all those pre-sales technical folk who do tons of work for customers for free) just want people to be aware of & take advantage of all the great technology that's being developed.
in my experience most of the problems that organisations encounter is when they try to mash together incompatible systems, under-spec their original requirements in order to cut costs, or simply don't understand what they're asking for.
Ceph, like Linux & others, is an open-source project.
RedHat bought Inktank (or, really, some of its engineers) & is now trying to figure out how to jam a Ceph product in with its Gluster & other storage stories.
SUSE's involvement with Ceph pre-dates RedHat's & it also has engineers working on the project, including some ex-Inktank. SUSE Enterprise Storage based on Ceph has been part of SUSE OpenStack Cloud distro for a few years, and as a standalone product for about 12 months.
"HANA was already available as an option on AWS for over a year, based on SAP's preferred & recommended platform, SUSE Enterprise Linux."
FIFY, especially since I couldn't find the Register article announcing SUSE's HANA on AWS availability way back when.
stripped-down, full-function enterprise linux?
welcome to SUSE, 2008
(although of course docker is just solaris c 2003)
let's also stop calling this "piracy" - which simply plays into the emotional argument framing people as blackhearted evildoers - and call it "unauthorised copying".
most of this copying /does not/ result in the downloaders making a profit by reselling - people are simply responding to the egregious pricing by old-world distributors who want to maintain gross revenue and rely on distorting markets through litigation and lobbying.
Artists have alteady been paid. It's the old-world distributors who are desperately trying to hold onto their outdated business model that want the extra $$$.
Press release, which the article didn't point to, is here: https://www.suse.com/company/press/2015/suse-enterprise-storage-powered-by-ceph-now-available.html
Right there it says 0.1c/gb/month for the software component.
Yes you need your own hardware: but the point is it can be *commodity* hardware, and a mix depending on what's best value at the time.
Pricing is based on a server starter pack + capacity model, but can be as low or lower than 0.1c /gb/month for the software component based on the fact you get 4x storage nodes in the basic pack. The higher density your storage servers, the lower the per-gb price.
just build what you want with susestudio.com ...
SUSE (note capitalisation) has been run as a separate business since the Attachmate acquisition of Novell.
It would make sense for MF to continue with that same approach.
The speculation about some kind of sudden integration of COBOL with SUSE Linux kind of misses the point of this kind of business.
When EMC acquired VMware it didn't suddenly force VM's to only run with EMC storage...
in terms of *private* cloud - i.e., changing your internal datacenter to use cloud concepts like policy-driven deployment, automation, templating and self-service, I'd argue a lot of enterprises _are_ lacking the operational maturity needed to fully realise the benefits (http://pwl-n-stuff.blogspot.com/2014/06/why-isn-cloud-taking-off.html). in particular, where the environments are outsourced, the big service providers have little incentive to make it easier to get things done with, e.g. self-service, if their financial model is based on getting lots of $ for performing additional but simple tasks that were left out of the outsource contract.
on the other side, though - where are the *enterprise* applications that support/require/capitalise on a cloud-type infrastructure? until they become more pervasive, there's less impetus for enterprises to adopt that kind of infrastructure & the complexity it entails. (http://pwl-n-stuff.blogspot.com/2014/07/cloud-high-availability-antifragility.html)
so they're offering samba-based windows domain integration & linux containers, have included XFS without charging extra like they used to, and have finally made it to a 3.x kernel. all stuff SUSE did years ago. congrats.
looks like they're also trailing on the btrfs front - although at least the tech preview means users will have a *chance* to include snapshots & so on, so long as they don't mind not having full support. i wonder if it's integrated into the system management tools yet, like it is in SUSE?
so who do you call for help when something goes wrong ?
so now you can get both enterprise linux distributions on Google Compute Engine.
except SUSE Linux Enterprise is, ah, less expensive for small workloads & large workloads. https://developers.google.com/compute/pricing#premiumoperatingsystems
The scalability discussion is no longer relevant.
HP's "Kraken" server is a single-system-image machine with up to 240cores and 12TB of RAM (24TB once bigger memory modules are available). It's designed to run SAP HANA, and that's developed on and only available on SUSE Linux.
On the insane side, SGI's Ultraviolet systems can scale with a single system image of 2048 cores and 64TB RAM. Again, the biggest of these systems have been run with SUSE Linux.
So, from a former Solaris OS ambassador: the scalability discussion is no longer relevant.
Vanilla OpenStack can be painful to set up, that's why SUSE Cloud also includes the Crowbar project mentioned in the article: basic infrastructure nodes (control, storage, etc) set up in ~20min; new servers can be just plugged into the network to be automagically discovered & allocated to roles.
OpenStack in general is targeted at public & private cloud infrastructure: last user survey had about 65% of respondents doing private. It also has the advantage of being hypervisor-agnostic: so you don't have to waste valuable vsphere licenses on basic stuff that can be satisfied with xen or kvm on linux.
this was always going to happen.
remember how this started out? Labor said "we're going to build this fabbo network", Abbot then immediately countered (without talking to anyone who had a clue) with "that's a bad plan, because, Labor bad".
from that point on, Turnbull was painted into a corner. I think it's pretty clear that to him, the concept of FTTP (implementation notwithstanding) was always going to be the preferable solution - but due to Australian politics he couldn't immediately come out & say that - he had to support the party line or else end up not being in any position to change it.
Instead he has been slowly slowly manoeuvring away from the original "NBN bad" towards one where universal broadband is acceptable. eventually once the coalition won government, he was able to implement "a review", which is often political-speak for "make it look like we're going to make sweeping changes, but instead take the heat off for a bit, while we work out a way to do the thing that we know is right, but that we have previously argued against for no good reason".
his recent comments hint that he's confident that they've found some technical loophole that allows fttp in a way that is "better" than the previous proposal - or at least a way to package it in enough gobbledigook to make it seem that way to the mainstream.
the next phase will be the announcement of the review's findings, indicating that there will be mostly fttp possible due to something special that the Labor govt "ignored", or something like that.
Australian politics as normal - at least Turnbull seems interested in the subject.
not requiring a physical sim means that you could easily have multiple operators for your phone without having to fiddle about swapping little bits of plastic. you could then select from a menu who you want to use at a given point in time, maybe have different networks for voice vs data, have one handset for work and personal numbers, etc. it also means you could order an international account *before* you arrive in a country & have it activate the moment you step off the plane (no need to find somethere that sells the right-sized piece of plastic to fit in your device).
also, if you have a subscription with a network operator, you should then be able to use any device simply by authenticating yourself: dropped your phone in the ocean? no problem, grab the old handset in your cupboard, reactivate it & your old number is still available to you. or add your own number to your friend's hadset for a while so people can still contact you on the trip back from the beach.
preporly executed this seems like a great idea (but then, you can say that about a lot of things...)
- "non geeks" use the functions of the computer they've been shown. windows or linux will be equally confusing to them, so why not use the one with lower licensing costs?
- in over 20 years of working life, i've yet to see *anyone* be sent on training for microsoft office, including the many iterations where the user interface changed radically. the "training" rationale is a furphy.
- if you've got a modern user productivity environment, then everyone should be using standard-based web apps, and the desktop OS is irrelevant. granted this is more of an aspiration than a reality, but people are slowed being dragged into the 20th century [sic].
what i find fascinating is that the customers who demand references are usually the ones who won't give references themselves....
MS office skills? really?
how much training actually happened when office 2007 came out & confused the hell out of everyone with its changed user interface? the swearing & cursing around the place went on for years, but no-one got trained. everyone's just expected to know.
here's the rub: if you're not spending $$$ on the software, you *can* afford to train people and they might (gasp) become more productive!
i hereby predict & coin the "IT reaggregation" trend of 2017-2020, where organisations "Precipitate" data and apps back from the cloud into central systems. the motivation will be partly security, but mostly for speed of service, efficient interaction of the widely-dispersed business apps & data, and improved control of the information environment for user groups frustrated with the huge number of disparate saas providers.
Microsoft is Mike Lee's equal 9th-highest campaign contributor
Herb Kohl isn't even standing next year, so who knows his angle.
why don't these providers find places where NBN is going to be years to roll-out & install stuff there? like down my street for instance :P