Scattered » Blog Archive » VMware Fusion 2 vs. Parallels Desktop 4: Let’s Dance

Home / asides / VMware Fusion 2 vs. Parallels Desktop 4: Let’s Dance

Glad to read this article after spending money on VMware Fusion 2. I had been a big parallels fan until I started using snapshots for IE testing.

(via VMware Fusion 2 vs. Parallels Desktop 4: Let’s Dance)

When it comes to OS virtualization on a Mac, there are two major contenders for the title of virtualizer to end all virtualizers.

Likely Parallels and VMware Fusion need no introduction for TAB readers, but you might not be aware of what the latest incarnations that both programs bring to the table. VMware Fusion 2, released in September, and Parallels Desktop 4.0, just released today, have a few new tricks up their sleeves.

Setup

Setting up both machines on my aluminum iMac was incredibly easy. I used Windows XP Media Center Edition from a physical disc for both, although the programs also offer the choice of using an image instead. For both installations I used the default settings. In Parallels 4.0, this consists of a 32 GB hard drive with 512 MB of RAM and 128 MB of video RAM. VMware’s quickstart configurations sets you up with 40 GB of disk space, 512MB of RAM, and although it doesn’t have a video memory slider like Parallels, 3D acceleration is enabled.
(more…)

Comments to “VMware Fusion 2 vs. Parallels Desktop 4: Let’s Dance”

Rob says:

November 12, 2008 at 9:37 pm

cool dude cool you just reblog, loser

dancameron says:

November 12, 2008 at 10:15 pm

Not exactly, these posts are generated from my Google Reader Shared items - that I comment on. I'm also blockquoting and linking back.

Reblogging (IMO) is strictly pulling in posts, without linking back and without any additional commentary or insight.

So…cool dude cool you just have ignorant comments.

JaredB says:

November 12, 2008 at 10:54 pm

Yeah, this is why I don't do that for when I pull my Google Reader shared stuff into my blog. Well, not *this* guy exactly, but potential complaints about reposting as well as potential dings from Google from their splog detector algorithms.

At first, I had it set up to copy the whole thing in, but then I changed it to only pull in whatever comments I make + the link itself. It's a tradeoff, because it is useful to have a copy in there in the event that the original goes away (which I've seen happen at least once since I started doing this).

I was just thinking the other day about pulling it in anyway and not showing it by default, putting it behind a no-follow link that would then show the content but not get caught up in Google.

dancameron says:

November 12, 2008 at 11:27 pm

no-follow is for links, but you could (or I could) setup these pages (based on category) to have a no-index tag in the header.

I think that's why I might do, even though I don't care too much about getting dinged from Google - anymore than I already am.

JaredB says:

November 13, 2008 at 7:28 am

? I'm pretty familiar with what no-follow does; that's why I said “behind a no-follow link”. If the content is only displayed when you click on that link (and you can't get there any other way), then that would work fine. There are other ways to tell it not to index certain content (headers, robots.txt, etc.) but those require more of a customization to your theme. The way I'm thinking of would not require theme changes.

JaredB says:

November 13, 2008 at 7:43 am

Oops, that last one didn't go in as a reply - probably because I was blocking the script at first.

Anyway, what are you using to pull the items in from GR. Is it still that feedwordpress (or whatever it's called) plugin, or something else?

dancameron says:

November 13, 2008 at 8:45 am

I'm using the feedwordpress plugin to pull in the content.

No-follow, would be pretty hard to accomplish because every link on your site, including archives, rss and plain listings on the homepage.

Modifying the template was easy, I updated the theme minutes after your comment, I just didn't post about it.
<?php wp_title(); ?></title>
<?php
if (in_category(3))
{
echo '<META NAME=”ROBOTS” CONTENT=”NOINDEX, NOFOLLOW”>';
}
?>
This way even the category pages have the no-index since they have excerpts. And when I create my archives that statement doesn't pass.

JaredB says:

November 13, 2008 at 6:30 pm

Sounds like that will work for you, but I'm wanting something slightly different.

I don't want the noindex header for the normal view of the items (which will just contain the links to the items and my comments, as they do now) for a couple reasons:

1. I still want the linked item to get the “link credit” for me linking to them

2. I still want Google to index those parts (the title and my comments), since I use Google for search on my blog rather than an internally based search system (because WP search generally sucks, but that's another subject).

Anyway, the way I will do mine won't require any theme modifications, which is valuable to me for a couple reasons. One is that I want to be able to change themes (or the basis of my theme) without having a core part of my content organization and display logic tied into the theme and having to duplicate that effort on any new theme I may want to move to. Another would be the ability to eventually release this system I build for pulling in these various items and having it have no dependencies on any particular theme (and not requiring any theme mods) so that others could use it more easily.

It will be pretty easy to do with a link, since this link will only appear in the post content itself, and it will be consistently reflected in all views (archives, feeds, etc.). Since the stored copy of the original feed content will not be a part of the actual post content, and will only ever be displayed after a user clicks on that link, it will be effectively hidden from search engines if they don't follow that link.

dancameron says:

November 13, 2008 at 6:56 pm

I still don't understand.

I'm guessing here:
Are you saying you would post the entire content but some of the content would be hidden behind a no-follow link?
If so, I think Google will see just about everything unless you're running an request to the db to pull the additional information. Because the robots may not follow the link but most robots still spider hidden content.

Or are you talking about posting the full content but having all of your links to that post with a no-follow?
Doesn't make sense to me, how would you put a no-follow on a title URL in WP without adjusting your template. And you'll be hiding the entire page from robots anyways since the robots couldn't find the content.

The real issue is duplicate content from another source and since SEs don't allow for wrapping content with a no-index, I'm not sure if there is any great resolution.

I'm probably going to take the no-index off the site, I'm just doubt the re-posting will be noticed because of the slight differences (i.e. my “note” and the blockquote).

JaredB says:

November 13, 2008 at 7:23 pm

Your first scenario is closer than the second. Here is a summary of how it will work:

These particular posts will appear on my blog just like they do now (fully available and indexable). The actual post content will contain (as it does now) a direct link to the shared item, a link to the source of the shared item (person or site whose feed I shared it from) and my comments / notes from GR, if any. See my GR category for how this currently looks / works, because that will all be the same.

The one addition will be a no-follow link in that post content (after my comments) to “show original content” (or something to that affect). Clicking this link will redraw that page, triggering a plugin that will load the original feed content (probably stored as postmeta data associated with this post) and insert it into the page for that load only (and any subsequent posts back to that same page, like adding a comment).

The primary reason I left out the original content in my first version (other than just consideration for people who may not like their stuff being re-posted) is the concern regarding duplicate content. Since the duplicated content concern is driven by the idea that it's not kosher to lure traffic (search results, etc.) to your own site by simply always republishing other people's work, I think indicating to the SE's that I don't mean for this content to be indexed (by making it inaccessible without following a no-follow link) should alleviate that issue.

In the end, these SE algorithms are a black box, so it may very well be true that they'll crawl it anyway, but from hearing guys like Matt Cutts talk, I think they're smart enough to recognize this method as a technique that's designed to avoid doing uncool things related to manipulating search rankings.

JaredB says:

November 13, 2008 at 8:23 pm

Maybe I should just go ahead and write this tonight as a proof of concept, and to avoid confusion.

JaredB says:

November 13, 2008 at 8:52 pm

Now, to derail this conversation back to the original topic…

I'm a big fan of VMWare, although it's unfortunate that the “Workstation” product is not available on the Mac. The huge feature that VMWare workstation has is what it calls “linked clones”. I believe this is actually exclusive to VMWare (probably due to patents) and the Workstation product specifically (as far as I know).

Snapshots are definitely nice for testing, but linked clones adds a whole other layer of benefits, especially on laptops, since they usually don't have enormous hard drives.

In case anyone is unfamiliar with the linked clone concept, and how it's different than the snapshot feature offered by most VM products, I'll briefly break down the differences:

Snapshots allow you to save a particular state of a given virtual machine installation, which you can then load and / or revert to at a future date. If you picture it as a timeline, you can install the guest OS and then set markers at points on the timeline that you can later revert back to, in the event that something goes wrong, or you just want to get back to that “clean” state.

What “linked clones” adds is the ability to “branch” off that timeline and have parallel lines that are all based off of a single underlying guest image, yet they can be independently updated, etc. without affecting the other branches.

An example is if you want to have XP VMs for testing using IE6 and IE7, but you want to allow both “copies” to keep up to date with the latest Windows Updates (other than updating IE on the IE6 VM of course). Since snapshot-only products don't allow you to make (and save) changes to a “snapshoted” state without screwing up future snapshots farther ahead in the timeline, you would need to make a full copy of the VM image and have two full-size VMs on your drive, one for 6 and one for 7.

With linked clones, you can start with your single base install (an up-to-date IE6 system, in this example), and then create a linked clone to a new VM and update that to IE 7. Now you effectively have two “branches” sharing the same underlying several-gigabyte XP image, and you can update and save changes to both branches as much as you need. Using this technique, you could theoretically have dozens of “different” XP installs on a fairly small hard drive, since each copy only takes up the disk space used by their differences.

dancameron says:

November 13, 2008 at 9:07 pm

The snapshot functionality isn't exactly a timeline, since I have snapshots of IE6, IE7 and IE8 with different dates and such. I'm able to move back and forth but I can't run them at the same time (not sure if linked clones allows for that).

Fusion does take an extra minute or two at most getting the snapshot ready for start up.

Do linked clones allow you to update the base install, then the branches will be updated? Seems crazy cool if so, right now I have to update each branch/snapshot if new OS updates come out.

And no, it's not copying the entire system for a new snapshot, only the changed files so it's fairly similar.

JaredB says:

November 13, 2008 at 10:36 pm

No, I think neither snapshots nor linked clones allow changing the base install with those changes being reflected in the linked clones or snapshots.

Perhaps they've made snapshots in Fusion behave more like linked clones in Workstation, but the snapshots I've seen do not allow you to save changes to “past” snapshots without screwing up future ones.

So, to clarify, if you have one base XP image, and 3 snapshots for each of the browser versions, and you made them sequentially (IE6 snapshot, update to IE7, snapshot, update to IE8), can you apply updates and save changes to the IE6 snapshot without trashing the IE7 & IE8 snapshots and redoing them?

If so, (and if the IE7 & IE8 snapshots are not full copies), then it sounds like they've incorporated the linked clone feature into Fusion's version of snapshots, although I don't think so. I think they would market that feature better, if it were there, since they tout that it heavily on their Workstation product pages and don't seem to mention it at all on the Fusion ones.

dancameron says:

November 13, 2008 at 10:58 pm

Yep, I have 3 snapshots and I can modify each one without screwing up another.

I just Google and found this,
http://stackoverflow.com/questions/61775/replac...
see the answers.

For those interested: I don't think Parallels has this type of functionality, the last version I used (v3) has a timeline type of feature.

JaredB says:

November 13, 2008 at 11:51 pm

Not sure if that's the one you meant to link to, but I did come across that article earlier, and it backs up what I believe to be the case:

You can only create Linked clones and Full Clones (which are close to differencing disks) in VMWare Workstation (not Fusion).

I did actually Google around a bit and found this and other posts that confirm my suspicions that linked clones are not available in Fusion, but if you're sure that you have them, you'll have to show me next time I come over.

Just to confirm, you're saying you have these snapshots that all shared the same base (no full copies) and you can open up an earlier snapshot (IE6, for example), add a bookmark (for example), and shut down the VM and have that change be saved. Then you can open a later snapshot (IE7 or 8), make changes to those and save them, then go back to the IE6 one and have that bookmark you added still be there?

JaredB says:

November 13, 2008 at 11:58 pm

Also see the last two comments on this thread on the VMWare support forums:
<a href=”http://communities.vmware.com/message/856208;jsessionid=FB1A124A5C2F51EE640F75F6D2EFB226″>http://communities.vmware.com/message/856208;js…

dancameron says:

November 14, 2008 at 12:42 am

So I made a video to prove Fusion is doing what I say it's doing.

I had to break it up into two files since the trial I have is limited to five minutes.

Quick run down (can't remember all of it):
Save current state
Open IE6, update Firefox and host file
Open IE7, save state (?)
Open IE6 original and show firefox is still not update and host file not changed.
deleted snapshots to show size change and saved states to see the space increase.

A lot in those 6 minutes but you'll get the message.

http://dancameron.org/wp-content/uploads/video/...
http://dancameron.org/wp-content/uploads/video/...

JaredB says:

November 14, 2008 at 8:28 am

Well, at least now I can see that we're talking about completely different things. Bottom line, I'm still pretty sure that Fusion does not support linked clones.

One thing you're illustrating here is suspend support, which is a whole (third) separate issue than linked clones and snapshots, and I think most VM products do allow for this. (Suspending the currently running machine - usually by storing the active guest memory in a file on a disk).

This is a different concept from snapshots and linked clones, though.

Snapshots will allow you to store a state of a given VM that you can always get back to. Depending on the product, you may also be able to go back and run the VM from the snapshoted state, and suspend it, etc., but NOT save changes to it that persist if you were to restart of shutdown the XP guest image (unless you get rid of snapshots after that point). You usually have to delete future snapshots if you want to actually save (persist across reboots) changes made to an earlier snapshot.

Linked clones would allow you to do the following simple test, based on two saved states (IE6 and IE7, for this example):
1. Open the IE6 VM. Create and save a text file on the desktop. *Shut down or restart XP*.
2. Open the IE7 state. Verify that the text file is NOT present on the desktop. *Shut down or restart XP*.
3. Go back and start the IE6 VM back up. Verify that the text file *IS* still present on the desktop.

If you can do this, while only having one underlying base XP image, then you have the “linked clone” feature. Maybe they just don't call it that in Fusion, but I can't image why they wouldn't, since they still call it that in Workstation.

If you can't, (as I suspect), then Fusion doesn't have the feature I'm referring to.

dancameron says:

November 14, 2008 at 8:36 am

So, just because I didn't restart the OS you don't believe the feature is there?

Okay I'll have to make another video, which is crazy.

When I'm done I want a big, “I was wrong”

dancameron says:

November 14, 2008 at 9:41 am

So I had to pause a lot in order to fit it in 5 minutes but I did this:

Save a IE 6 snapshot
Save a file in IE6
Open IE7 and the host file wasnt updated
Go back to IE6 with saved file and verify it's still there

You'll notice that both systems are different now, one has far cry 2 on it.

Also, not on the video. I'm able to go back to the IE6 without the saved file (pre-snapshot save) and do whatever I want with it, without affecting the other snapshot.

http://dancameron.org/wp-content/uploads/video/...
So?!?!?

JaredB says:

November 14, 2008 at 12:51 pm

Don't worry, I don't mind saying I was wrong if it turns out I am. I'll even put it in all caps, to make it as big as possible.

JaredB says:

November 14, 2008 at 12:59 pm

Interesting, but it's still not quite clear. Just a few questions:

1. You're positive that those various XP snapshots are all small, change-only copies on the disk, right?

2. I think there is still some fishy stuff going on with suspending and creating those extra snapshots.

What I'm really curious about is the 3 step scenario I described above (using *just two* “snapshots” and not creating new ones in between), and *never suspending* but doing a full VM shutdown in between the steps (and not creating additional snapshots).

You don't have to make a new video if you're tired of it though.

It's not that I don't trust you, I just think we're not quite on the same page in terms of understanding the nature of the feature we're talking about, and since all the info I've been able to find says that Fusion doesn't have linked clones, I'm hesitant to believe that it does.

dancameron says:

November 14, 2008 at 2:00 pm

1. Yes, they're all the off the same. In the video they're all attached to “XP”. In the videos yesterday, I showed how creating the additional snapshots wasn't copying the entire base.

2. Huh? I can turn them off or suspend them, I just prefer suspending.

3 step: That's exactly what I did, except I created a new one to show that you can create an additional snapshot.

Finally, we're on the same page you just don't believe that a Fusion for –mac– is as good as workstation.

dancameron says:

November 14, 2008 at 2:01 pm

Also, snapshots probably wasn't supported in Fusion 1.0 and you're reading old content.

dancameron says:

November 14, 2008 at 2:25 pm

Actually, we're back at square one because I don't think you understand what I've stated Fusion can do.

In order to keep the changes you must save the “current state” in order to start a different snapshot.

So in step one I would need to create a copy, otherwise that state will just be deleted before I start up IE7.

But you're question was, “can you apply updates and save changes to the IE6 snapshot without trashing the IE7 & IE8 snapshots and redoing them? “
The answer is yes, you apply all the updates you want to IE6, then save it, then open up IE7.

Basically, if you want to keep your work you need to save it.

Maybe it's not exactly linked clones but I never said it was, I just answered or tried to correct you.

For example,
“Linked clones would allow you to do the following simple test, based on two saved states (IE6 and IE7, for this example):
1. Open the IE6 VM. Create and save a text file on the desktop. *Shut down or restart XP*.
2. Open the IE7 state. Verify that the text file is NOT present on the desktop. *Shut down or restart XP*.
3. Go back and start the IE6 VM back up. Verify that the text file *IS* still present on the desktop.”
Yes, you can do this in Fusion. You just need to save the state.

Also,
“Snapshots will allow you to store a state of a given VM that you can always get back to. Depending on the product, you may also be able to go back and run the VM from the snapshoted state, and suspend it, etc., but NOT save changes to it that persist if you were to restart of shutdown the XP guest image (unless you get rid of snapshots after that point). You usually have to delete future snapshots if you want to actually save (persist across reboots) changes made to an earlier snapshot.”
It's not a timeline, you don't trash other snapshots - because it's not a timeline.

You're essentially creating independent snapshots using a base install but you need to save the snapshots because they're not independent VMs.

JaredB says:

November 14, 2008 at 4:59 pm

No, I was ignoring most pre 2.0 stuff. I'm primarily just going off their product pages themselves, where they mention and highlight the linked clones feature for workstation and do not list it for Fusion.

Any link that states otherwise would go a long way towards convincing me; the more official the better.

JaredB says:

November 14, 2008 at 5:04 pm

I think we're still not on the same page, actually.

I believe you with #1 - just checking.

Re: #2 - My point there was that IE6 keeping the file could be explained by the saved suspend state. Turning off is definitely different than suspending, because suspending stores a large file on disk containing the guest memory state. Just trying to rule out that variable in order to prove your point.

We can probably just put this on hold until I get a chance to see it for myself (or see a VMWare link that confirms it), because until then I will probably continue to have doubts that we're both talking about exactly the same thing.

JaredB says:

November 14, 2008 at 5:10 pm

I agree that we're not talking about exactly the same thing. I'm strictly talking about the “linked clone” feature as named by VMWare and included in their workstation product.

Let's just leave it at that, for now, since it seems to be too difficult to figure out going back and forth here in the comments.

Next time I see you we can come to a conclusion, (unless you can post a link before that), at which point I will happily and publicly admit to being wrong if that's the case.

Then I'll complain about VMWare's web pages on multiple sites for being extremely unhelpful and misleading about this particular issue, because they definitely currently make it sound like Fusion doesn't have linked clones.

JaredB says:

November 14, 2008 at 5:17 pm

Back to this side conversation, (and on the subject of me being wrong), it appears that we're both right/wrong on the nofollow issue.

You are right in that the “official” intent of the nofollow attribute for links is only related to page rank calculations, not indexing, so the name is actually a bit confusing in that respect, because it doesn't actually mean “don't follow”, but rather “don't give credit / juice / etc.”

That being said, Google and others also do not index the content of a page that's arrived at via a nofollow link, which is what I had suspected, so I believe my original plan will still work. However, I'll probably take other measures to prevent indexing (only when the hidden content is shown) to try to keep that hidden content from being indexed by the other minor SE players, while still allowing the stuff I show now (title, link and my comments) to be indexed since they will be included in the default view.

Setup

Post Details:

Related Entries: