Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Ubuntu Switches To OpenStack For Cloud

timothy posted more than 3 years ago | from the disintermediating-reintermediation dept.

Cloud 55

angry tapir writes "Canonical has switched its cloud software stack to the open-source OpenStack. The current version of its Ubuntu Server, version 11.04, uses the Eucalyptus platform. Ubuntu Server 11.10 will include the OpenStack stack as the core of the company's Ubuntu Enterprise Cloud (UEC) package. The server release will also include a set of tools to help users move their cloud deployments from Eucalyptus to OpenStack."

cancel ×

55 comments

Sorry! There are no comments related to the filter you selected.

DON'T RAIN ON MY PARADE !! (-1)

Anonymous Coward | more than 3 years ago | (#36127892)

I love openstack for cloud. It makes all the sense of a summer's eve. Another dochebag name !!

Ubuntu (-1, Offtopic)

Osgeld (1900440) | more than 3 years ago | (#36127992)

stable as a (crack) rock

deploy your servers with us, you will get so good at moving shit, you wont need a baby sitter distro much longer

Re:Ubuntu (0)

Anonymous Coward | more than 3 years ago | (#36128014)

Am I to infer that crack rock is unstable? In what way?

Re:Ubuntu (2)

at_slashdot (674436) | more than 3 years ago | (#36128066)

You are using of course LTS for your servers, right? Because otherwise it would be stupid... So, is 3 years support not enough for you?

Re:Ubuntu (1)

Osgeld (1900440) | more than 3 years ago | (#36128070)

what's Microsoft's server support?

Re:Ubuntu (4, Informative)

the linux geek (799780) | more than 3 years ago | (#36128114)

Typically about ten years (Server 2000 support ended last year, and 2003 is still supported.)

Re:Ubuntu (1)

froggymana (1896008) | more than 3 years ago | (#36131160)

Microsoft doesn't offer *free* upgrades to their next version of the their server OS though either...

Re:Ubuntu (0)

Anonymous Coward | more than 3 years ago | (#36128122)

LTS is actually 5 years.

I think MS's is longer, but not by orders of magnitude. NT4 was EOL around 2004, so 8 years. Server 2003 ended regular support last year (so 7 years), extended support will be available until 2015 (so 12 years). I'm not sure what to compare to what.

All that said, the LTS does not mean that they actually include more stable versions of software, so I would also be wary of using LTS on production servers, as 10.04 LTS was as stable in april 2010 as 11.04 is now.

Re:Ubuntu (1)

laurelraven (1539557) | more than 3 years ago | (#36135092)

10.04 LTS was as stable in april 2010 as 11.04 is now.

Sure; however, for those using the LTS on servers, if they were on version 8.04 LTS server, they have until April 2013 to upgrade (when they drop support for 8.04). This means they could wait until 12.04 LTS to have a full year of stability patches before upgrading; meanwhile, they have a server that is very stable in the sense that their software packages continue to work the way they did when they first installed (no updates to newer versions, just patches to improve security/stability).

"Stable", in server terms, doesn't just mean "doesn't crash"...just as importantly, "stable" means "doesn't change". Give an LTS a year to mature, and it will give you both. Buy a server box to put it on, and by the time it has given you the four remaining years on the LTS, you might very well be looking to replace the server with something more powerful or with a fresher warranty anyway. Then, when you set up it's replacement, you skip an LTS for the latest and test to make sure it all works on the new version. Congratulations: other than regular security patches, you won't have to do much else with that box for another 4 years (assuming you have it set up the way you need initially, and that your needs for it don't change much).

Re:Ubuntu (0)

Anonymous Coward | more than 3 years ago | (#36128082)

LTS is 5 years for servers [ubuntu.com] .

Overlap between one (4, Informative)

tepples (727027) | more than 3 years ago | (#36128170)

LTS is 5 years for servers

About two and a half years of that overlap between one version and the next. It takes about six months to get a .04.1 respin that's stable enough to offer as an upgrade to users on the LTS channel. There's even less overlap if you're trying to use the same LTS version on your development workstation/server and your production server. Let's take Ubuntu 8.04 (Hardy Heron), the operating system currently offered on Go Daddy dedicated servers running Plesk, as an example:

  • 2008 April: Ubuntu 8.04 released; five-year period starts
  • 2008 October: Ubuntu 8.04.1 released and officially blessed as LTS
  • 2010 April: Ubuntu 10.04 released
  • 2010 October: Ubuntu 10.04.1 released and officially blessed as LTS
  • 2011 April: Support for desktop components of Ubuntu 8.04 LTS ends
  • 2013 April: Support for server components of Ubuntu 8.04 LTS ends

Re:Ubuntu (0)

Anonymous Coward | more than 3 years ago | (#36128154)

LTS releases are supported for 5 years (server).

Re:Ubuntu (3, Informative)

polaris20 (893532) | more than 3 years ago | (#36128220)

You are using of course LTS for your servers, right? Because otherwise it would be stupid... So, is 3 years support not enough for you?

Actually, Ubuntu LTS server is five years. http://en.wikipedia.org/wiki/Ubuntu_(operating_system) [wikipedia.org] In fact, 6.06LTS server support is just wrapping up this June. Desktop versions of LTS are three years.

Re:Ubuntu (1)

kesuki (321456) | more than 3 years ago | (#36128086)

to be fair i have tested many software and found them all to have issues, somewhere for someone.

if you have stable enough hardware any infested or otherwise software will still run.

Hmmm (0)

Anonymous Coward | more than 3 years ago | (#36128112)

Is OpenCloud from the goatse guy?

pieces of an incomplete cloud solution (1)

rubycodez (864176) | more than 3 years ago | (#36128126)

so where's the distributed database system to go with this solution, that scales to thousands of nodes and billions of records in tens of thousands of either tables or hierarchical structures like xml, yaml or whatever?

Re:pieces of an incomplete cloud solution (1)

snelders (1021195) | more than 3 years ago | (#36129588)

I think it doesn't exist yet (true scalable distributed ACID compliant database cluster). So far I think NimbusDB is the most promising project trying to realize such a thing.

Re:pieces of an incomplete cloud solution (1)

Compaqt (1758360) | more than 3 years ago | (#36132422)

Well, mongodb [mongodb.org] is being used by a lot of "cloudy" web 2.0 companies (Shutterfly, Foursquare, Bitly, and others [mongodb.org] ).

It doesn't use XML or YAML, but you can build hierarchical structures with JSON because it's basically a free-form database.

What sort of application would need thousands of nodes (other than Facebook)?

It does automatic sharding (horizontal partitioning), with eventual consistency.

Re:pieces of an incomplete cloud solution (1)

rubycodez (864176) | more than 3 years ago | (#36133670)

yes, there are huge scalable databases out there, I'm saying OpenStack should have one. That upper limit (thousands of nodes) was just an example, a venture that is planning to be big would want something with no scaling limits. Even if they only wind up with dozens of nodes, investors will want to see that they planned for a big future.

Complexity kills reliability (4, Informative)

gweihir (88907) | more than 3 years ago | (#36128172)

Leslie Lamport' s comment on distributed systems applies:

"A distributed system is one in which I cannot get something done because a machine I've never heard of is down."

This is even more so with the "Cloud". Think 99.99% uptime? Then why had some Amazon customers recently to wait 5 days to get access to their data again only to find out it was not all there? Don't get me wrong, cloud computing has its place, for example short term high-CPU or high-bandwidth needs. It can be used as a redundant (secondary, _not_ primary) system for e.g. Web-Servers. It is also nice, if you can rent a high-memory instance when you have the occasional (rare) job that needs more memory than your own machines have. Also virtualization has its place, namely as a HAL on steroids.

One thing the "Cloud" is not usable for at all is high-reliable server services. Another is processing of any confidential data. It is not self-redundant either, there are single points-of-failure, as Amazon recently demonstrated. For redundant, reliable infrastructure, you have to do your own primary systems, the "Cloud" can at best serve as fail-over. These limitations do apply to private clouds as well. For longer-term high-CPU needs, your own infrastructure is far, far cheaper and better tailored to your needs. For processing anything confidential or secret, public clouds are unusable and private ones need the whole private cloud classified to the highest secrecy level processed on them. You may also have to have several ones of each classification level if there is a horizontal isolation need (i.e. you may not process secrets from A with secrets from B). At some point the cloud becomes a problem, not a solution.

Why everybody is driven to the "Cloud" like lemmings is beyond me. It is one more tool with specific limitations and strengths. It is not a one-size-fits-all at all.

Re:Complexity kills reliability (-1, Troll)

ArsonSmith (13997) | more than 3 years ago | (#36128212)

Spoken like a true buggy whip manufacturer. Well done.

Re:Complexity kills reliability (1)

gweihir (88907) | more than 3 years ago | (#36128232)

What are you talking about?

Re:Complexity kills reliability (0)

Anonymous Coward | more than 3 years ago | (#36128274)

Whip that buggy dammit! Lol.

Re:Complexity kills reliability (1)

sourcerror (1718066) | more than 3 years ago | (#36128394)

It's buggy for sure.

Is it good enough? (4, Insightful)

js_sebastian (946118) | more than 3 years ago | (#36128236)

Leslie Lamport' s comment on distributed systems applies:

"A distributed system is one in which I cannot get something done because a machine I've never heard of is down."

This is even more so with the "Cloud". Think 99.99% uptime?

(In many cases) The question is not whether the cloud gets you 99.99% uptime. It is whether it gets you better up-time than what you can run in-house for the same price. It's easy to insult the amazon guys when they fuck up, but the availability they offer is certainly better than what a small company can get from their single part-time admin who does something else as a day job. And even if you are a small tech company, where in theory anybody has the knowledge to run a few services, in practice it is very easy to make mistakes, even for smart people.

And when you scale up, the cloud can scale up with you. Of course, by the time you're google you'll be running your own data centers...

Re:Is it good enough? (2, Interesting)

Anonymous Coward | more than 3 years ago | (#36128888)

Do you have anything to back those claims up? Anything at all? What you describe is not what we're finding in the real world, when dealing with real businesses and real "cloud" platforms.

Over the past few years, over 30 of my business clients moved to some sort of a "cloud" platform. Of those, about 20 have admitted that they were wrong and have moved back to more traditional hosting for their web applications and infrastructure. This has mainly been due to reliability issues, and also because these "cloud" providers turned out to be far, far more expensive than running their own servers in-house. Of the remainder, several of them went out of business due to their "cloud" platforms making them uncompetitive, and a few more are seriously considering a move away from their current "cloud" platforms.

You may think that I dislike cloud computing, but I don't. I'm very glad that it has arisen. Fixing all of the problems it has caused has been a very profitable venture for me! Not only that, but it has also helped make my clients realize how much value there is in dealing with real servers that are in their possession, or at worst at a nearby data center.

Cloud computing is a fad, and its time is already almost over. It has failed to live up to the hype, and it has failed to live up to just about every promise made about it. Businesses are abandoning it faster than ever.

Re:Is it good enough? (0)

Anonymous Coward | more than 3 years ago | (#36130198)

"Do you have anything to back those claims up? Anything at all? What you describe is not what we're finding in the real world, when dealing with real businesses and real "cloud" platforms."

You mean like Google?

Re:Is it good enough? (0)

Anonymous Coward | more than 3 years ago | (#36131206)

A cluster you control isn't a cloud, at least not in the original sense (the one he is obviously using). Hell, the term comes from the depiction of the internet on diagrams, cloud computing is (was?) by definition letting someone 'out there' take care of the gritty details and only dealing with the abstraction.

Re:Is it good enough? (1)

LordLucless (582312) | more than 3 years ago | (#36132466)

He's got about as much to back it up as do you and the OP. Namely, opinions and anecdotes.

Re:Is it good enough? (1)

Ender_Stonebender (60900) | more than 3 years ago | (#36132924)

"Cloud computing" is this decade's "The Network Is the Computer". (Remember that?) It got slightly more traction because the network has actually considerably improved since the late 1990s, but the problems are essentially the same. I suspect we'll get another round of this bullshit, under a new name, sometime around 2024.

Re:Is it good enough? (1)

syousef (465911) | more than 3 years ago | (#36132136)

Leslie Lamport' s comment on distributed systems applies:

"A distributed system is one in which I cannot get something done because a machine I've never heard of is down."

This is even more so with the "Cloud". Think 99.99% uptime?

(In many cases) The question is not whether the cloud gets you 99.99% uptime. It is whether it gets you better up-time than what you can run in-house for the same price. It's easy to insult the amazon guys when they fuck up, but the availability they offer is certainly better than what a small company can get from their single part-time admin who does something else as a day job. And even if you are a small tech company, where in theory anybody has the knowledge to run a few services, in practice it is very easy to make mistakes, even for smart people.

So what you're arguing is that cloud computing makes sense not for large businesses which demand reliability but for small ma and pa operations too poor or too cheap to hire a sysadmin. Only trouble is cloud computing isn't targetted so much at small business, but more at the top end of town which may demand 5 9s or 6 9s.

And when you scale up, the cloud can scale up with you. Of course, by the time you're google you'll be running your own data centers...

But you just said it can't be expected to if you want high uptime.

Re:Is it good enough? (1)

laurelraven (1539557) | more than 3 years ago | (#36135268)

...the top end of town which may demand 5 9s or 6 9s.

6 9s? You mean, service that is SLA'd to be down no more than 31 seconds per year? Is it even possible to promise that?

Not trying to troll here...I'm serious: is that actually a usable measure?

Re:Complexity kills reliability (1)

Anonymous Coward | more than 3 years ago | (#36128244)

You must not have heard about Xeround. It can fail over to multiple cloud vendors and in fact it remained available during the re-mirroring storm at Amazon even though they had instances running on the degraded datacenter.

Private infrastructure is not cheaper. The maintenance overhead is very expensive when rolling your own infrastructure.

Re:Complexity kills reliability (1)

outsider007 (115534) | more than 3 years ago | (#36132358)

Thanks for the info, Shilly McShillington

Re:Complexity kills reliability (1)

QuantumRiff (120817) | more than 3 years ago | (#36128532)

Another is processing of any confidential data.

Why not? Why is an amazon EC2 instance, with encrypted data and IPTables only allowing port 22 and 80 any less secure than having the same thing, sitting in the server room. I have never understood the security fears of "the cloud" except from people that think security is putting a firewall at the network edge, and trusting anything that is behind it. If you deal with defense layers, and you monitor your systems, the location really doesn't matter as much.. (except for physical security, of course)

Re:Complexity kills reliability (2)

gweihir (88907) | more than 3 years ago | (#36128934)

No security against the cloud provider is the main issue. Also any security audit would have to include the cloud software as well, in the version that is going to be used. For higher classification levels, it is often illegal to put the data on hardware that also processes non-classified data or data from other organizations. With EC2 you have no control what other data will be on the same hardware. It is still forbidden, for example, to process credit card information on EC2 and with good reason.

You can also not process encrypted data in any meaningful way today. Fully homomorphic encryption is still very much a dream, although with a recent (eminently non-practical) breakthrough by IBM Research, it may not be impossible after all.

Re:Complexity kills reliability (0)

Anonymous Coward | more than 3 years ago | (#36129070)

PROCESSING, not TRANSFER.

Cloud is transparent to the person who provides the service, or at least that's usually the case. And there always will be at least that one engineer who wants to take a peek at the interesting stuff that may be stored there.

Security is NOT a problem with The Cloud! (0)

Anonymous Coward | more than 3 years ago | (#36128776)

Wait a minute. I'm a manager, and I've been reading a lot of case studies and watching a lot of webcasts about The Cloud. Based on all of this glorious marketing literature, I, as a manager, have absolutely no reason to doubt the safety of any data put in The Cloud.

The case studies all use words like "secure", "MD5", "RSS feeds" and "encryption" to describe the security of The Cloud. I don't know about you, but that sounds damn secure to me! Some Clouds even use SSL and HTTP. That's rock solid in my book.

And don't forget that you have to use Web Services to access The Cloud. Nothing is more secure than SOA and Web Services, with the exception of perhaps SaaS. But I think that Cloud Services 2.0 will combine the tiers into an MVC-compliant stack that uses SaaS to increase the security and partitioning of the data.

My main concern isn't with the security of The Cloud, but rather with getting my Indian team to learn all about it so we can deploy some first-generation The Cloud applications and Web Services to provide the ultimate platform upon which we can layer our business intelligence and reporting, because there are still a few verticals that we need to leverage before we can move to The Cloud 2.0.

Re:Complexity kills reliability (3, Insightful)

MichaelSmith (789609) | more than 3 years ago | (#36129762)

Cloud computing services are ideal for situations where you have a startup which might fail in two months and you don't want to have to install a warehouse full of computers to get it going.

Why everybody is driven to the "Cloud" (1)

doperative (1958782) | more than 3 years ago | (#36133482)

> Why everybody is driven to the "Cloud" like lemmings is beyond me

Because the venders hope to make more money then selling the one server.

Re:Complexity kills reliability (0)

Anonymous Coward | more than 3 years ago | (#36139128)

Our philosophy with cloud and outsourcing is that if its something generic that can be pushed outside of our organization we want to look at doing that and put our in house resources towards things that are unique to our organization.

You are talking about edge cases where cloud may not be a good fit. For general usage however a cloud provider can give far better security, uptime and resources than the normal company can manage. Most of us do not run server farms of over 600 servers with redundant facilities with redundant power able to go longer than 24 hours and multiple high performance network backbones coming straight in with admins with a skill level to match.

There were providers such as Netflix who use Amazon's cloud who were unaffected because they did failure planning. People mistakenly thought that the cloud was unfailable and they got a reminder. If you still do your needed due diligence you can be fine, its not a silver bullet but its not something to put your head in the sand over either.

Ubuntu is adding, not switching, to OpenStack (5, Informative)

martinbogo (468553) | more than 3 years ago | (#36128760)

Hi. I'm one of the ARM Server developers who just attended UDS Budapest. In fact, I'm still here at the hotel.

Ubuntu did not _switch_ to OpenStack. Rather, Ubuntu has added OpenStack as another method of creating a personal Cloud using Ubuntu. By doing so, we're adding to the rich diversity available in the Ubuntu universe. It's not replacing Eucalyptus! Eucalyptus remains supported.

-Martin B
ARM Server Developer
(In Budapest)

Re:Ubuntu is adding, not switching, to OpenStack (0)

Anonymous Coward | more than 3 years ago | (#36129778)

Yes. It's OpenStack _and_ Eucalyptus, not OpenStack _instead_of_ Eucalyptus. Not as sexy a headline.

Re:Ubuntu is adding, not switching, to OpenStack (0)

Anonymous Coward | more than 3 years ago | (#36130664)

I knew something had to be up, since it wasn't making headlines on my other news aggregators.

Re:Ubuntu is adding, not switching, to OpenStack (0)

Anonymous Coward | more than 3 years ago | (#36130762)

> It's not replacing Eucalyptus! Eucalyptus remains supported.

And once OpenStack is fully deployed and up to speed, how long after that will Eucalyptus remain supported? If you are Canonical, what's the business case for doubling your cloud dev costs in this area once the new one is capable of running everything on its own?

just sayin'.

(see also the history of SourceForge)

Re:Ubuntu is adding, not switching, to OpenStack (2)

slashdot_commentator (444053) | more than 3 years ago | (#36131136)

1) Eucalyptus will be supported for as long as Ubuntu versions are supported. So, for 10.04 LTS server, that will be until 2015. (barring some catastrophic dissolution of the Eucalyptus company.)

2) Eucalyptus is not dependent upon Ubuntu for its existence. Amazon is currently the largest cloud vendor using eucalyptus tools to interface with its services. If Amazon drops eucalyptus (ha ha), then you can start the death clock. You're not going to be able to get openstack to nicely merge stuff with EC2, so its not likely openstack will supplant eucalyptus anytime soon.

3) Canonical has no announced plans to remove eucalyptus from its distributions. Most likely, eucalyptus will be in 12.04, and that means it'll be around for at least 2017.

Ubuntu on OpenStack (0)

Anonymous Coward | more than 3 years ago | (#36128846)

Freaking awesome.. I work for Rackspace and didn't know that THIS was coming down.. ! :)
Anyone in here from the Ubuntu project involved want to comment on how it we came together on this? I know that we have a LOT of Ubuntu lovers at Rackspace. We even have our own internal mail list at Rackspace called ubuntu@rackspace.com (not external obviously). :)

My favorite distro and my favorite cloud storage solution... two great tastes..

Tweeks

Re:Ubuntu on OpenStack (1)

tomweeks (148410) | more than 3 years ago | (#36128896)

[repost authenticated]
Freaking awesome.. I work for Rackspace and didn't know that THIS was coming down.. ! :)
Anyone in here from the Ubuntu project involved want to comment on how it we came together on this? I know that we have a LOT of Ubuntu lovers at Rackspace. We even have our own internal mail list at Rackspace called ubuntu@rackspace.com (not external obviously). :)

My favorite distro and my favorite cloud storage solution... two great tastes..

Tweeks

Story link may be dead (1)

slashdot_commentator (444053) | more than 3 years ago | (#36130910)

You can try this one [desktoplinux.com] instead.

Go at Canonical (1)

CondeZer0 (158969) | more than 3 years ago | (#36131836)

Interestingly at Canonical they are starting to use Go for their backend infrastructure [cat-v.org] .

I wonder if they will start to replace components of the grid stack with stuff written in Go like Doozer [github.com] .

Re:Go at Canonical (1)

pmontra (738736) | more than 3 years ago | (#36132052)

Oh my, a language with pointers. I thought they were recognized as a worst practice and forbidden in any modern language. It would be nice if they started playing with this other go [wikipedia.org] instead.

Re:Go at Canonical (1)

CondeZer0 (158969) | more than 3 years ago | (#36132442)

Before displaying your silly prejudices it would be useful if you informed yourself a bit.

Go has pointers but no pointer arithmetic, which allows it to be safe unlike C. Also, Java and pretty much every other 'modern' language has pointers, all objects are passed by reference, but the programmer has no real control over the memory layout of structures and pointers are 'hidden' from the programmer (most of the time) and you are left at the mercy of the design decisions the creators of the language made; this is one of the many reasons why Java sucks so much [cat-v.org] .

What Go provides you is explicit control over memory layout and whatever you pass things by reference or by value, this is extremely useful for systems programming, and gives the programmer much more clear control over what his code is doing.

Re:Go at Canonical (1)

pmontra (738736) | more than 3 years ago | (#36132714)

I know that Go has no pointer arithmetic and I'm fine with 'modern' languages's pass-by-reference and taking control of memory layout. Somebody even believes that this might lead to better performances, for the same reasons compilers might be better at optimizing programs than we are. However I understand that in some cases you have to know exactly which byte goes where or you want to fit as many data as you can in a small amount of memory. I did that in C many years ago and I understand the need to do that but let me quote Fran Allen [wikipedia.org] , from Chapter 13 of Coders at Work [codersatwork.com] . She's more qualified than me to make this point.

Allen: By 1960, we had a long list of amazing languages: Lisp, APL, Fortran, COBOL, Algol 60. These are higher-level than C. We have seriously regressed, since C developed. C has destroyed our ability to advance the state of the art in automatic optimization, automatic parallelization, automatic mapping of a high-level language to the machine. This is one of the reasons compilers are . . . basically not taught much anymore in the colleges and universities.

Seibel: But most newer languages these days are higher-level than C.Things like Java and C# and Python and Ruby.

Allen: But they still overspecify. The core thing is that it specifies location of data. If you look at these other languages, they stayed away from specifying the location of data and how to move it, where to put it in the machine. It was ultimately about its value at any point.

[...]

Seibel: But very few languages other than C and C++ have raw pointers anymore. Java has garbage collection and the data moves around. Would you say that’s still overspecified?

Allen: Yes. I believe that there’s an opportunity to do what we have done with computation in the optimization world with data. We don’t manage data very well. We don’t have good ways of managing data automatically — establishing locality of data that’s going to be used together. There are lots of threads of research now which are very exciting. But I think what’s missing is the bigger, bolder concepts. A lot of this is happening within a space that is bounded by what exists already or the current thinking. It’s not going to change overnight by any means—there are millions of lines of code out there. But we do need to start trying to break the boundaries of, “This’ll be done here and that’ll be done there.”

Seibel: Can you give a simple example of what you mean by bringing the data to the computation in contrast to what we know how to do now?

Allen: To me it means taking over the management of the data, Basically, how we do it now is by reference—it’s moved by hardware, or by the underlying operating systems and support systems. [...] But another way to do that would be to organize locations of data in their relative positions as a target of optimization. The other part of it is that very often what is good for one computation is poor for another. One organization, even of simple things like matrices, is bad when you’re actually accessing it in a different way. So it’s a combination of the order of the accessing against the location. It may require some architectural work, and hardware work, but I think that one can do this if we put some of the referencing, addressing capabilities back out in the hardware itself. There are machines where one has the ability, at the point data comes into the memory, to do quite a lot of transformations. Mapping can happen there. Computation speed is what we measure, mostly, in high-performance computing so we go through all kinds of things to increase that speed. Feeding that computational unit is one of the big issues that we face, but we never made it a first-order problem to solve. We leave it to the hardware.

Seibel: In your Turning Award lecture you said something along the lines of, “We’re at a crossroads here and we might miss it. We might go down the wrong path and then be going down it for quite a while.”

Allen: Yeah.

C and Go are glorified assemblers. If you have to do a task which needs an assembler they are the right tools and Go might be better than C (I never tried it). Pointers have no need to exist anymore outside those specialized tasks.

Peace.

Check for New Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>