745 posts • joined Tuesday 19th June 2007 21:44 GMT
It's mostly because they can -- you get SIGNIFICANTLY more out of a 3PAR system than you can an EVA system. Massive consolidation is possible.
Per unit parts may cost more, but you should end up saving quite a bit(or at least break even) because the # of those parts you have to buy to serve a given workload is significantly less.
And for the AC posting - note that 3PAR has had iSCSI support going back at least 7 years now, in the past year or so they added 10GbE iSCSI+FCoE.
when did xyratex get out of the disk array business? they still make enclosures for sure, HP uses em(on 3PAR 7000 at least), last I saw IBM used them on V7000, NetApp stopped using them I believe after they absorbed that division from LSI. I imagine a bunch of other folks use Xyratex still too, not sure about HDS/EMC I don't make a point to track Xyratex :)
looks like Xyratex still makes their own arrays for HPC as well.
just the title of the article shows how clueless the author is (never said that about an el reg author before). amazon has been absurdly priced for MANY MANY YEARS now, again it comes down to the architectural flaws in how the system was designed(they couldn't do a better job at the time because their technology sucked, and they're so far down that suck road now that it's really really difficult to change gears at this stage).
But wait.. there's more. Google has the same model.. HP cloud is the same model.. Rack space Openstack cloud is the same model!! Openstack in general (last I was briefed on it anyway) adopts that same broken model (again because the technology isn't good enough to do it right). The likes of VMware, Hyper-V, RHEV etc got the model right, it's just complicated to scale to very high levels. So these public cloud players take the short cut, of massive compromise on features, availability, and capacity utilization in exchange for massive scale == large costs for the VAST VAST majority of applications.
You're basically reversing a efficiency trend in IT that has been going on for the past 8-10 years and going back to 90s era provisioning strategies, at least from an IaaS perspective. PaaS, and SaaS are provisioned and billed with different models and those can make sense at large scale operations for customers(obviously both of those have their own limitations). the IaaS technology the public cloud folks are making available is just absolutely terrible by contrast.
But this author doesn't know what he's talking about, I don't know why I bother to keep writing comments in his articles I suppose I should just stop.
(El Reg reader for the past ~14 years)
how is google more stable?
they will be just as unstable as any other public cloud provider.
They may seem more stable now because there aren't as many users.
At least they provide VM fail over (haven't looked into the details). still far too many missing things in all of these clouds to make them worth using for any serious work.
that doesn't allow pooling of resources is a giant fail in my book, no need to continue discussion. that includes the likes of amazon, rackspace, hp, etc..
we've been pooling resources in the "enterprise" for nearly a decade now, going back to 90s-era infrastructure is just wrong in so many ways.
exactly - these numbers are coming from a market space that Apple doesn't even bother to compete in. Apple's CEO has specifically said they are not going after that market. They are perfectly happy where they are at..rolling in more cash than most anyone else combined in the mobile space.
If apple ever does feel threatened they can easily drop the prices as a last resort, their cash stockpile is so big they can sustain just about anything for the next 10-20 years.
(I say this as a WebOS user -- still waiting for galaxy note 3 64GB to show it's face before I make the jump to android.)
PC Power & Cooling hangs onto their stuff.. they always seemed to make quality stuff in my experience, and I noticed a couple of years ago they were bought by OCZ.
Every discussion I've seen around OCZ has always centered on their SSDs.
those 10,000 servers dropbox uses to front end it's services.. even though they store the data in s3.
doesn't make sense to me if you invest in 10,000 servers that you don't take the time to invest in the object storage as well, but whatever.
Re: Don't be fooled...
your certainly going to scare away any good candidates, the only ones left will be the ones who can't find a job elsewhere, who have been stuck in jobs where they are exploited. I knew a couple guys like that in the past. It was really sad to see how they always ended up at really shitty gigs. They had opportunities at times but never took advantage of them so they remained at shitty gigs.
of course the devil is in the details maybe the company only has 5 people in it or something and the IT side of things is trivial, in which case most things are outsourced to various SaaS companies anyway. Maybe the extent of their IT is managing a printer.
you lost me at cloud
but I'll pay shipping if they want to send me 10TB worth of disk drives.
More like probably 6th or 7th? Where do you get this second biggest from? Networking is cut throat, you have to fight tooth and nails to get market share. Arista from a tech standpoint may look good compared to Cisco - I mean shit, how hard is it NOT to look good against a 10+ year old switching platform.
Obviously the people that are using the 6500 1) don't care about costs 2) don't care about performance(they'll just buy more) 3) don't care about power efficiency/usability/etc 4) don't want to invest in anything new(e.g. change) - they are happy with the old way of doing things.
I remember reading up on marketing stuff way back in 2004 on chassis switching platforms that destroyed the 6500 in almost every way (except sales of course). I'm sure there was stuff earlier but that was about the time I started getting more involved in networking.
Just looking at when Arista was founded- Oct 2004, these PDFs I have came out before Arista was even founded(and obviously years before they shipped any product)
Maybe those previously mentioned cloud players(another article) that are "shunning Cisco" were dumb enough to do wide scale deployments of 6500 and that's why their costs are so high.
Re: whiny little bitches
I think it's fair to assume with exponential growth comes exponential costs, there's only so far down you can drive stuff. You can bend the cost curve a bit, but if your already purchasing probably $10s-100s of millions worth of networking gear per year you already have a lot of the benefits of such scale already.
Now maybe Cisco is able to gouge these cloud players as much as they can the little guy, I don't know, in which case they probably only have themselves to blame as there have been good alternatives to Cisco going back 15 years now.
Even using IOS makes me want to gouge my eyes out with an ice pik, it's so needlessly complicated.
whiny little bitches
(disclaimer - I've never been a Cisco customer, and never will be)
Those big cloud players have enough purchasing power to likely get Cisco gear at the same cut-rate price as the competition offers to other players. I remember one big Cisco deal many years ago where the customer got 90% off Cisco list pricing just to keep the competition out (Cisco had otherwise lost the deal on all fronts whether it was performance, cost, etc).
I recall again many years ago(5+) that Amazon had been building/deploying their own custom "TOR" switches with I believe broadcom chipsets. probably goes back further than that. Though big core switches were probably from one of the big players, same goes for probably google and others too. I don't track networking too closely but have not heard of any large scale "white box" switches - e.g. that do hundreds of 10Gbit ports at line rate speeds for aggregation/core workloads.
Look at the average cost of a server today vs 10 years ago and likely you'll find it quite a bit more expensive. ASP is up across the board on everything I think, though what you get for that price is also significantly more than it used to be.
I remember when a 4-8 port 10Gbps blade for a core switch would run easily $40-50k+ (and to even consider 10Gbps at the time you had to have a chassis switch) -- now we get 1U buggers with 1Tbps of fabric and 56 ports for under $20k, or 48 ports with some minor restrictions for well under $10k - in almost all cases that little 1U 10GbE switch has significantly more fabric capacity than the 10-20U+ core switches of the earlier era), and obviously a small fraction of the power draw.
What you used to pay a quarter million for on a core switch to do ~100-500Gbps of switching fabric you can easily eclipse 10Tbps today.
So, yeah. A bunch of whiny little (or I guess big) bitches.
don't forget 3par 10k
HP doesn't even talk about the HDS stuff any more, haven't heard a peep about it in over a year now. You likely won't see it mentioned in even marketing literature (at least I haven't seen it mentioned again in more than a year).
They still sell it (though I'd be shocked if they hopped onto whatever is the next generation beyond VSP), but they would rather customers buy 3PAR of course.
Maybe that Suresh guy is busy updating his resume.
reminder for microsoft
Your domain onmicrosoft.com expires in Feb 2014, not far away, you should renew your domains for a longer period of time. Even my own personal domains were extended to 2016/2017 several years ago.
So don't forget microsoft, renew that domain before it expires. Wouldn't want a repeat of that hotmail(?) problem you had a few years back!
all 17 organizations out there that use VDI will be impressed.
Re: Budget Lumia, or this?
motorola mobility was even wiser to sell itself to google
why it took em so long to join the party
why they aren't on a common storage platform (software at least) across their major product lines
When they will post SPC-1 results for XtremIO
Don't worry, I won't be holding my breath.
Re: Real competition
This is EMC, they don't care about technology, they care about marketing etc. I'm as shocked as anyone else that they've managed to not only hold their own against startups over the past decade but grow their market share with the collection of mostly crappy solutions that they have (there are a few exceptions).
I don't believe this will change substantially with this round of startups, but I suppose anything is possible
done nothing but cause problems
The marketing hype behind cloud is pretty good, because it's somehow got most non technical idiots convinced it can save them everything. I was talking to one such idiot recently who was interviewing to be my new boss(I'd be shocked if he got the job), I had to physically hold myself from bursting out into laughter on several of his comments/responses about "cloud".
But more seriously, the Amazon cloud had been BY FAR the biggest fail I've ever seen, BY FAR the biggest frustration in my professional career, nothing else is even in the same galaxy cluster as this thing. Other clouds are similar from a cost perspective at least (as in cost is so bad it's a "face wall" moment). Companies spending literally a quarter million to half million or more in cloud fees PER MONTH are quite common in my experience. Spend a day learning what such companies do and you can quickly realize how much waste is going on.
But the waste itself is not a problem in itself - the problem is the quality that these cloud services provide is universally very poor. I would feel differently if the quality was really high, but I've yet to see(or hear) of a provider where the quality was really good(regardless of cost - I have been told stories of even the most expensive cloud players out there having terrible service/performance/etc).
I talk to folks fairly regularly these days now about them either 1) wanting to move out of a public cloud 2) are moving out or 3) already have moved out. Yet these *rarely* make the news, it's always the inverse.
My two most absurd reasons for using cloud that I have heard of in the past year:
"I don't want to have to deal with servers, or vendors" - from a company spending $600,000/mo in a cloud (not Amazon). They understand they can build their own stuff for around $2.5M - but it doesn't matter. The shitty cloud they use(they have quite a few outages) is what the management likes.
"I think(hope) we can attract (more) investors if we are in a cloud" - from a sinking company who wants to move to Amazon, they acknowledge/understand that public cloud will cost them minimum 3-5x what they are paying now. Yet they hope that is offset by additional investment from VC(s).
Top reasons for moving out of a public cloud:
- Unpredictable performance
- Bad/non existent support
I know one company that gets in excess of 80% off list pricing for a major public cloud and even with that pricing they still want to get out ASAP as they've had endless technical problems using the public cloud, some of which cause more than a day of lost productivity for each glitch.
For me the technical(and business) reasons go FAR FAR BEYOND the above top reasons, but the marketing has been so successfully drilled into these clueless management types that it's like trench warfare getting them to understand that cloud is bullshit.
Fortunately for my own sanity I have not had to deal with public cloud shit in well over a year now. The ~3 years I had working with amazon gave me PTSD, I start twitching when I hear that word.
SO THANKS FOR THE PTSD
With this firefox plugin I can now have a more positive experience when reading about cloud. This article for example comes across as
"Ask the Sysadmins: What has my butt ever done for me?"
one word: awesome (said in the voice of Cartman)
good use case
this is one of the VERY VERY FEW good use cases that is good for amazon (spin up a bunch of shit for 18 hours then kill it). Unfortunately the # of folks that need something like that are tiny. So Amazon markets to the general hosting audience for which it is horribly suited.
Another similar good use case for security folks is password cracking esp with GPU acceleration. I suspect most of the folks that do this probably will do it with stolen CCs and stuff because that's just what they do, most "legit" folks don't need to crack passwords. I think this use case came up a few years ago.
I installed this plugin on my firefox a short time ago and it made this article much more enjoyable to read
Last I read Backblaze was quite strict in which drive models they use(I think for the most part hitachi drives).
Re: The year of VDI?
Also the year of linux on the desktop.
Just wanted to say THANKS again to Ed Snowden for doing a great job in opening the world's eyes to more of what goes on in secret.
Damage or not, I'm amongst those who believe firmly that it is worth it.
SO THANKS FROM ME.
Think about it - if Netflix has only what 25-30 million subscribers and they already account for this much of the internet's traffic? Doesn't seem like a scalable system to me.
I am still old school - while I haven't rented a video from a video rental store in more than 20 years(my mother's boyfriend was personal friends with the original owner of Blockbuster and that guy used to send me a bunch of free rental tickets back in ~1991), there was a "recent" introduction of premium channels like HBO, Showtime, Movie channel, cinemax etc where I have gotten the bulk of my movies over the past 15 years at least (combined with Tivo which makes for easier watching of course - I haven't been without a Tivo on my personal TVs in 13 years now).
One of the side effects of watching everything through tivo though is I rarely am exposed to what is new, often times I find out about cool tv series long after it was canceled, or I flat out don't hear about a good movie, or if I do I forget about it by the time it hits the movie channels (sometimes if it sounds cool enough I set a Tivo title+actor wishlist to grab it if it won't be showing for another several months).
I did do Netflix for a while but after a few months quickly ran out of content, and then for the next year found I didn't watch more than 1-2 hours for an entire year (not per month but combined for the whole year) - (Netflix used to email me "how was the quality on X?" not sure if they still do that - I went back through those emails and realized that almost all of the titles in the previous year it asked about I did not watch more than a few minutes of) - I canceled when they raised their prices, I was renting an average of 1 dvd from them every 1-2 months(there was a bunch of dvds I wanted to see but I was just lazy when it came to returning them). Not really enough to be worth while. I keep hearing about content complaints and it seems their whole licensing model is broken (not their fault I guess since they can only do what the studios are willing to do), so stuff can come and go at any time.
I do watch a bit of youtube from time to time, not too much though.
The fabric controller started spewing gibberish?
"if a top of rack switch fails you've got another one - redundancy is built in," he said."
But if the controller starts spewing gibberish across the control plane there goes your entire cloud in a giant POOF. Probably makes STP outages look like a kids movie by comparison.
Re: I dont get it...
yes you don't. I don't know what Oracle 12c offers but the multi tenant features I can think of all revolve around resource utilization. Sharing I/O, sharing CPU, sharing buffer cache, sharing DB connections etc. Don't want one user to be able to adversely affect another. With Oracle I think back to all of the outages I was a part of that were related to latch contention, that's a good way to kill all other tenants in a Oracle DB pretty quickly.
IMO has nothing to do with basic level security.
That doesn't stop most folks from running multi tenant databases and perhaps most of the times that strategy works fine (it has in my case for years on MySQL). But there are times when one app blows up and it causes issues with others on the DB.
Virtualization would be a more ideal way to go about it, since things are more partitioned, but obviously much less efficient from a hardware utilization standpoint.
probably doesn't matter
The sorts of attacks the NSA and similar agencies are likely to launch are going to be quite sophisticated, and not likely to trigger detection by regular AV systems. They'll be highly targeted. So even if the software ends up on systems other than the target the payload won't be executed, because they want to keep the attack secret, it doesn't help the NSA if the AV firms are out there (and there will always be one or two that will not comply even if the F-secure folks are lying - though I don't think they are since they don't have to).
There may be accidents, like Stuxnet getting out(even though the payload was not executed outside of Iran last I heard it still caused some other issues). Though such attackers have probably learned from that and are being more careful now.
Too much to write here, wrote on this topic a couple months ago. I know el reg doesn't like links to other sites but I can't duplicate the writing here with images etc..
basically sdn is a crock of shit. I see it useful for the hyper scale players out there, but the number of organizations where SDN will be really useful is quite limited. Network vendors haven't gotten the attention that the storage and server virtualization folks have gotten over the years. Networking is viewed as boring, it's basically a utility -- how often do you see people getting excited about a new UPS or power strip..? There have been some interesting things happening at layer 7 over the past decade but that's layer 7.
So networking companies have over hyped this SDN concept to hell and back to try to make things sound exciting again. A few years ago it was FCoE (and DCB/DCE) -- obviously that flopped!
Now it's SDN..
When in reality -- the people that really need this stuff (and I admit there are customers that do) -- already know about it. In a lot of cases I bet (Google, Amazon etc) already have a been doing a sort of SDN long before it was formalized as a term, because they got to that scale where they had to do something.
But in the link above I address more directly the flaws I find in the hype of SDN in general and how it's really just that - hype.
I did not feel comfortable attacking SDN dead on before recently because until recently I could not get an informed opinion of WHAT THE HELL SDN IS. I had the opportunity to ask the question to the inventor of SDN himself personally, which confirmed all my original thoughts and expectations and allowed me to write a good blog post on the topic.
I feel sorry for the networking companies I really do. I do wish they would spend more time on ease of use rather than adding ever more complexity. I was quite disappointed when I learned a few years back that TRILL was a layer 2 only protocol (I fully expected it to have a layer 3 component, because you know, we do this thing called ROUTING now and need redundant routers). I've really liked the protocol ESRP from Extreme which combines layer 2+3 (you can also do layer 2 OR layer 3). And Extreme in general has stuff that is easy to use, with a config language that more or less reads like english (configure vlan my_vlan add ports 1,2,3 etc).
But I suppose day to day ease of use is not flashy, not a catchy thing because you can just hire yourself an expensive network engineer and/or enroll in some complex training classes to learn how to use the equipment.
Look how easy (for the most part) modern storage is today(certainly are exceptions), or modern server virtualization. Most any idiot can fire up a vmware system and build a cluster and create vms with a few mouse clicks. sometimes that causes problems, but it's still easy to use.
Now some things in networking can be complex still, take routing protocols and stuff like that, that's fine. But basic L2 and L3 stuff should be dead simple easy to understand and build.
Perhaps in 3-5 years, maybe more, SDN will be integrated to the point where it's in a similar state. I don't know. If that's the case it's sad it took the industry 10-15+ years to get to a state of ease of use.
I have no doubt a decent part of the problem is the network admins themselves who pride themselves with their advanced certification and training courses on the overly complicated network equipment and they consistently over design things and just make life in general more complicated.
give me a dislike button
wtf is all this BS about like, where is the dislike? As a pessimist I demand a dislike button on all these shitty social media sites!
I mean even el reg has a thumbs up and thumbs down, you guys get it..
While we're at it I want to be able to vote "no" when voting in political elections (e.g. cancel out someone else's vote FOR a particular candidate).
Wonder how many thumbs down I will get for this.
can still be simple
As simple as you want it to be....it only gets complicated if you make it complicated.
I know one really big customer that for some reason is deploying 3PAR storage as a Commvault backup target (?!?!@!$#@ what? do they not know that HP has StoreOnce ?) and they use HP P9500 for their tier 1, but they don't use any snapshots(too complicated for charge back apparently).
Storage has been complicated for a while, I think back to my early days with thin provisioning on 3PAR with the company saying "oh go throw out your volume managers!" -- and learning the hard way long before thin reclamation was possible to do data migrations to reclaim space from apps that aren't efficient. Or the days where you had active/passive controllers on your array and you had to worry about shit like "LUN tresspassing", or some crappy storage systems where performance went in the tank once you started turning on some of the features(still quite common today I hear).
Perhaps storage is getting SIMPLER - by having more silos, at least for larger orgs (not me) - having more single/special purpose systems(unix way of "do one thing and do it well") instead of systems that try to be everything(and do so poorly).
best security defense
IMO the best security defense you can have is not to be an attractive target to begin with. Obviously not possible for many big organizations.. But for smaller ones, that will get you more than any technology or training or policy.
It usually comes down to the value of your data to other folks, if your data is not that valuable (reality here not what value you put on it personally) then it's less likely to be attacked, there's other more tempting targets.
That doesn't protect you against worms/viruses which don't discriminate on who they attack of course but those things can often be effectively blocked by traditional security measures IDS/IPS/firewall/AV/etc etc..
Perimeter defense almost worthless
Saw an enlightening talk by the CTO of Trend Micro a few months back, hoping the org that hosted the event posts the video online, been waiting months and nothing yet.. here is another video with him from another event that is similar (though the talk I saw was ~45min this is ~10)
I didn't learn anything but it was awesome to see the honesty come from someone in his position where he just comes out and says it - if someone wants to get in, they get in. Just accept it. Doesn't matter how much you invest it won't be enough. His summary of the RSA break in was entertaining as well. He admits his own industry is at fault in putting a false sense of security in their customers "just buy this product it will make you safe". At the end he was pitching a new product of theirs but he said on several occasions that it won't protect you against everything.
Here's a PDF from the presentation I was at
Security is not my focus so I don't try to stay on top of everything in that field but his coverage of the bank attack in Korea was quite good too (page 36 of the PDF), nearly 50,000 computers disabled. In all 76 tailor-made malware were used, targeting both Windows and Linux/Unix systems.
I normally skip keynotes and stuff at events(the events that I do attend, which in itself is rare). This guy was just great though.
Perhaps the most ..informative stat(?) he cited is in 2012 the average time an attacker has access to a network before detection is 210 days (35 days longer than 2011).
Seeing that makes me glad security is not my focus, because really those guys are fighting a war they simply cannot win. "Ooh look this new shiny firewall or IDS/IPS!! But we still got hacked......."
I dealt with a hacked system on Sunday, first time in probably 5 years I've been involved in one. System wasn't being managed by anyone, ran a wordpress blog that was out of date. It seems like some worm got in by one of the recent code execution exploits, for some reason it wiped out the data on the wordpress blog itself (seems careless because that's how it was detected almost immediately) and downloaded some files and basically turned the host into a bitcoin mining operation(complete with fake program signatures to make it look like normal processes were running, and a crontab to re-download itself every so often). I don't believe they ever got root as they never wiped any of the logs, the apache error logs clearly showed the exploit in action. The IT folks restored the system from an earlier backup and updated wordpress. Now they want to transition responsibility of this system to me but I want nothing of it. It was an interesting exercise though, took only maybe 5 minutes to determine what was going on, obviously not a sophisticated attack.
In the event on Sunday there was absolutely no fancy security system in place, the incident was picked up on basic monitors within minutes of it happening because the worm/attacker wiped out the data causing the website to return errors.
who's gonna buy them?
IBM just bought TMS, EMC bought extreme, HP bought 3PAR, Dell bought compellent, NetApp bought LSI and has their flash ray thingie coming.. Cisco bought whiptail.
I think Cisco would be at least somewhat worried about pissing off NetApp/EMC even more (and thus losing some of their steam in their efforts to leverage N/E to get into accounts etc) by competing more directly with them. Perhaps if/when UCS growth tails off they will get to a point where they don't care and just go full steam into storage too.
The rest of the folks seem to be pretty happy with the various companies they acquired and probably are not likely to make further acquisitions anytime soon, IMO of course.
HDS, and perhaps Fujitsu(?) seem to be some of the only major folks left that haven't bought anyone recently ?(I don't count BlueArc since that is just a gateway). Anyone left out there?
Oracle I guess, Pillar is a door stop so I don't count them. They also have ZFS which I've seen almost nothing but complaints from customers on how poorly those systems are built/bad support etc. Though Oracle is all about the engineered systems, and putting the storage intelligence in the application layer where it's more efficient (Exadata), so I suspect not a high likelihood of Oracle going after these startups, unless it's *really* cheap.
I suppose there is a chance one of the cloud service providers may see value in some of this tech and acquire it (with the intention to use it entirely as a service provider role not re-selling appliances to customers). Though to-date it seems the cloud players are all about building stuff in house, no matter how crappy it may end up being.
Might we see something new here? Might some of these flash startups merge with each other, or acquire each other? That would be sort of interesting.
savings of up to 50% over 3-5 years ?? how about savings of 50% within the first year and that's not even bothering to use shitbox equipment with the complexities of open stack. That's tier 1 enterprise hardware and software with 24x7x4 hour response times from the vendors. No need to re-design your application for a built to fail architecture.
At least that's the story I've seen time and time again. One guy I talked to just last week was at a company that was at one point paying $500,000/mo to joyent, and they later brought things in house for closer to (avg) of ~$50k/mo.
I've done the same with amazon, and done math for others (one that was paying $300k/mo to EC2 the ROI there was about 4 months) - the cost is rarely the issue, it's normally bullshit politics and braindead CxOs and managers who see shiny cloud and think it will save their ass instead it will sink their budget. I talked to another guy recently who's company wants to move their shit to amazon because they think it will make them look more attractive to investors (they are aware it will cost them 3-5x+ more, but they will be in "cloud" and that is shiny).
Hoping the pending implosion of the current tech bubble will shake some realities into how much waste is going on with public cloud spending. It's just absolutely staggering. It's one thing to pay a lot for a high quality service, but the service provided by the public clouds is about as basic as you can get. It's depressing.
I'm sure there are some exceptions....in all the companies I've worked at, and all of the friends I have at other companies(and the VARs I work with who have customers using cloud services) .. I have yet to come across any *personally*.
all I need to know
about open stack is Red hat still hasn't manged to get a stable release of their product out yet. Also it appears they gave up on their original plans to go with Folsum, and now are working on Grizzly. The way it seemed to be pitched to potential customers last year I thought they'd have a non beta product 6+ months ago. But checking again today it's still beta, for non production, no pricing and no obvious release date noted for a non beta release. That tells me a lot right there. The other point is Rackspace (earlier in the year) abandoning their original willingness to support "any" open stack system out there.
Not that I care too much either way, I'm not in the market for anything open stack related. But it is interesting to see all of the hype around it, along with articles like this(and others) that talk about the realities of the current state of open stack.
vmware no longer exciting
IMO their last big good release was 4.0 (still using it for my production systems, well 4.1U3 - and I use ESX not ESXi - I like the thick mgmt stuff). I looked at what was new in 5.0, 5.1 and now I think 5.5, and don't see much that gets me really excited. 4.0 is obviously pretty old at this point, but the point is the competition is/has caught up to vsphere 4.0 in many areas.
So there's less incentive to go hog wild with vmware, though I still think it's the best hypervisor out there(and the only one I use today still).
Never been much into conferences myself, living just a dozen or so miles from VMworld was hosted I never had any interest at all in attending. My boss went for one day on a vendor invite - though he said he didn't see much either, given the lack of news on el reg on the thing I wasn't too surprised.
As a VMware customer for the past 14 years now (back when vmware was linux only, need to find my 1.0.2 CD ..) - I feel that it is no longer exciting, it's a very useful technology to have and I use it every day, but it's pretty mature tech at this point.
I told our VMware rep I don't see us buying any more licenses until at least 2015, I plan to upgrade the CPUs on our existing clusters, the cost of that is a fraction of additional hosts, and doesn't consume any additional power.
After that perhaps late 2014 start replacing existing servers with the newer faster variety.
The latest Intel 12-core chips are 3x faster than our existing 12-core 2.3ghz opterons, so we can triple our performance/capacity at some point and won't have to worry about any more vsphere licensing. Too bad that AMD abandoned the high end chips, was sad to learn that. Not sure what chip perf will look like a year from now.
Also I'm a strong believer in not making everything a VM if it doesn't have to be. If you can drive performance on the HW side (or there is a software licensing issue) I have no problems putting something on real HW (even though right now all of my shit is VM).
Re: Why Android?
most linux apps aren't touch oriented would be one.. and I assume you've heard of Windows RT ? You know, the Windows 8 that is a port to ARM ?
Why Android, well I think the answer is pretty simple - it was there first, and has a ton of momentum, that is very hard to counter, second is cost for the folks making the phone. Making a (good) mobile OS is expensive and difficult, making an ecosystem around it even more so. I've never used Android for more than a few mins though I am planning to go to Android from WebOS at some point soon.
I do feel it's a sad state of affairs that Android has basically become the Windows of the mobile era, I suppose it could be worse though (speaking as someone who uses Linux exclusively on their desk/laptops).
7450 can have 240 SSDs / 96TB on the data sheet (for a 4-node system)
Though I think the data sheet may be obsolete given they have 600G ssds now, which would make it 144T raw.
Re: Wait for manufactured speed numbers
SPEC does validate performance/latency/efficiency numbers along with cost with a common workload across vendors so for that I think it provides a good value. Obviously users shouldn't base their decision entirely on those numbers alone but they can provide insight into trends in the system to see if it's worth pursuing.
So for example if the cost of the system is really abnormally high relative to competitors a customer can determine that it is not worth spending the time to engage for that platform if they don't have the $$ for it anyway.
Or if the platform gets good performance, but nearly hits the 47% unused utilization ratios of the SPC-1 benchmark then maybe they decide they'll go for another platform that lets them use more of the capacity and still get good performance.
Or they can look at the full disclosure to see the system configuration, and determine that HEY this thing is really complicated and I want something easier to manage.
I remember one vendor from China had disabled their read cache prefetching in the array and devoted everything to write cache for SPC-1. They also happened to disable write cache mirroring as well (also
disclosed in the full disclosure - no mention of it in the shorter executive summary). They managed to get several hundred thousand IOPS - so impressive headline numbers, but the devil was really in the details.
SPC-1 is also a lot more trustworthy than the numbers that most vendors spit out themselves. e.g. lots of grand claims about 1 million IOPS.
I remember one time one of the 3 letter storage vendors came in to my company and quoted really high I/O numbers (abnormally high for the platform in question) - only after I challenged them did they confirm that was purely from cache.
SPEC sfs is less valuable to me at least since there is less disclosure, especially on the cost side.
Re: not enough
mainly the freakishly frequent release cycles that seem to constantly break things in an effort perhaps to stem losses to Chrome. I've never used Chrome for more than a few mins myself, but have seen tons of comments from users that have switched back and forth and the impression I get is if the users want something released often they go to Chrome(instead of wanting firefox to be more like Chrome - I haven't really seen anyone make the comment that they want Firefox more like Chrome -- they just go use Chrome), there's another set of users who just want something stable.
There is the firefox extended release thing, but apparently that has it's own flaws (which I read up on why Ubuntu does not use Firefox ER).
I suppose I could switch to something like Opera, not sure how often they break stuff, though been so used to the firefox plugins and how things work it would be a difficult switch. My Ubuntu 10.04 LTS systems are currently entombed on Firefox 20.0, I don't see that changing any time soon, I've taken reasonable steps to mitigate security risks. Before Ubuntu 10.04 LTS desktop edition went end of support the firefox updates were coming quite frequently and breaking things on a regular basis. There are still some things that are broken in firefox 20 but at least I don't feel concerned about new things breaking.
but not practical for most folks. Web scale obviously doesn't apply to most orgs. Which is one of my biggest beefs that is the crap of some public clouds like amazon. Most folks don't understand it's built to fail. They don't build their apps in that manor. Very few do. I've worked at a bunch of places including two who launched their production from day 1 in amazon cloud, none of them built their apps to such standards. Even after having massive issues in cloud as a result of built to fail they still don't change their models. Look at how many web sites go down when amazon has a hiccup. It's a wide spread issue.
It makes sense at big scale - though most orgs will never get there.
built to fail is just a whole lot of fail for I'd wager 99.9% of organizations. Really the only way I see this changing in a big way is if the built to fail model is further abstracted away by the platforms. So in salesforce case, the interfaces they expose to the end users/developers writing stuff for sales force may be quite reliable because the heavy lifting is done by salesforce itself. For those writing their own apps at a lower level though, not using a PaaS or something I don't see built to fail growing gaining a lot of traction. Organizations would rather write code that brings in more users then write code that is more robust to failure.
One of the folks I know had a pretty priceless quote recently "this code is so bad, that i need to go to the ocean to make my tears look small"
just more stuff that can break
All that technology sounds fancy, but to me it just sounds like your making things just a whole lot more complicated, and with complication comes more bugs, across more layers, and more risk.
For me just keep it simple - separate the data from the VM, replicate data between data centers, don't replicate the VM. Have another VM(or means to build one(s) fairly easily which benefits you regardless) in the other facility with the other ip address space and the right default gateway etc. If resource utilization is a big issue then just make the VM in the other facility really small until it comes time to need it and then crank up the CPU/memory knobs.
I guess maybe I am just more old school, I still run my server network links active/passive. My fiber channel links are active active though(round robin) - though there's really no extra complexity with active active vs active passive from a FC MPIO stand point in my experience on 3PAR anyway.
sounds like it works then
"Or at least, it doesn't play nicely alongside Unity, which I wasn't able to use again until I had un-installed Cinnamon 2.0."
Sounds like something Ubuntu should adopt!
I'm still on Ubuntu 10.04 LTS desktop even though it's out of support mainly because I'm so used to the GNOME 2 interface it has. I have no interest/reason to change, though eventually(I've decided that time will be when I get new hardware that doesn't work with 10.04) I'll have to upgrade to something and Mint or Ubuntu with Cinnamon sounds like the next best thing.
Firefox may not be getting updates anymore on 10.04 but I suppose the upside to that is it's stopped breaking shit every few weeks as a result. I run firefox under sudo (to access another user account not for root) under a 'firefox' account that has limited rights(which makes some things complicated as I have to trade files between accounts in /tmp), and java stays off most of the time, along with flash click to run, security wise just being on Linux I don't feel there is a huge threat to me.
Sort of reminds me of the HP Pre3, a bunch of folks said similar things about that phone (which I use daily) when it finally got into the hands of users even though it was canceled. They said *this* was the phone that should of been the launched a long time ago - not the Veer, maybe not even the Pre2 etc etc..
I was always *absolutely* convinced the Pre3 wouldn't make a dent in anything especially having a launch date so close to an Apple product, the idea was suicide really(turned out to be literally I suppose). WebOS needed a ton more work. I did feel that at the time(before the axe fell) that Pre3 would just be another step in the journey.
BBOS sounds like it works alright, just missing the good selection of native apps. I don't know BB like I know WebOS - but I suppose it's not hard to guess the Z30 isn't going to change BB's fate.
I'd be pretty happy with just a good web browser, photo viewer, and video player along with basic messaging and phone calls. Unfortunately even the Pre3 is faltering for me in most of those categories these days.
Hoping a 64G Note 3 comes out and that turns out to be a good device for me, never really used Android (or IOS) before - but I do like the idea of having a lot of knobs and buttons to use to customize, one of the things a lot of users seem to hate about Android.
I thought about Windows Phone, and BB as possible platforms but I think my use of Linux as my desktop means Android is likely going to be the best bet. I don't see a future for Ubuntu phone or firefoxOS either (Canonical+Mozilla have been consistently pissing off their user bases for the past couple years at least, can't imagine them any more successful in mobile).
wonder what their ratio is now
"Apple captured "19 percent of all sales dollars" of consumer electronics sales in the U.S. during the holiday quarter of 2011 according to NPD."
anyone know what the answer is now? I think I might be scared to know the truth.