Opinion: RHEL 5 turns 3, Suggestions for Red Hat

Red Hat Enterprise Linux 5 (Tikanga) was released on March 14, 2007 and yesterday was RHEL 5's 3rd birthday. Since then we have gotten 4 update releases.

Given the fact that Red Hat's original plan was to have a new RHEL release every 18 - 24 months, one has to wonder where RHEL 6 is and why it is so late. My best guess is that RHEL 6 (which so far has had a non-public alpha release within Red Hat as witnessed in some Bugzilla reports) will come out sometime this summer... possibly in time for the Red Hat Summit in Boston (June 22-25, 2010). For that to happen I would expect a public beta for RHEL 6 to be released in the not too distant future. We'll see how that pans out.

While we are waiting, how about some idle discussion?

What do people want in a RHEL 6?

RHEL 6 is said to be based on Fedora 12 so I hope to see both newer server apps as well as desktop apps... and of course a much newer kernel (2.6.32).

It appears that Drupal 7 requires PHP 5.2 or higher so it won't even run on a stock RHEL 5. Lots of newer Python-based web apps want a newer Python than is in RHEL 5. On the server side, people will be very happy to see RHEL 6.

On the desktop side we hope to get KDE 4.something, a much newer OpenOffice.org, Firefox 3.5 and a ton of other update apps. RHEL 5 is just ancient on the desktop.

Suggestions for Red Hat

In an article I wrote some time ago I had several suggestions for Red Hat. It seems that many of them are still worth repeating today... with some updates.

Previous Suggestions Rehash

Email / Messaging Product - We still need a good email server. Originally I recommended Red Hat buy Zimbra. Unfortunately that isn't an option any more since Zimbra was purchased from Yahoo by VMware. We still need a good, as free as possible, email server. People seem to get confused when I bring this up and point me to various SMTP servers. That is NOT what I'm talking about. I'm talking about a more complete solution that includes spam filtering, virus checking, shared calendar, shared documents, to-do lists, global address book, a modern / attractive web-based portal, and possibly even instant messaging... all rolled into an integrated, easily manageable system that can scale to meet the needs of any size business. You know, like Zimbra. Perhaps an option is Zarafa... or some FOSS project I'm less familiar with.

Additional Office Apps - Another suggestion from my original article was for Red Hat to buy SoftMaker and release SoftMaker Office as a FOSS product. OpenOffice.org is certainly nice and mature but it could stand to do with some competition. I've looked a little more closely at SoftMaker Office's credits and it appears to be based on a lot of third-party technology not owned by SoftMaker. I don't know why I didn't notice before but yeah, that would be a big challenge.

SoftMaker has been constantly improving SoftMaker Office and the 2010 version for Linux is expected real soon now. Their pricing and licensing, even though it is a completely closed product, seems somewhat reasonable for what it is. They do often give older releases away and this past holiday season they actually donated money to charity for every download of SoftMaker Office 2008 for Windows or Linux that they were giving away free of charge.

It would be wonderful to see SoftMaker Office open up, and hey, they offer more than 10,000 fonts too (probably also a license mindfield). I wonder how far fetched a web-based version of SoftMaker Office is?

Of course I'd love to see the KDE folks get somewhere with KOffice but I'm definitely not going to hold my breath. The GNOME office apps just aren't integrated well enough so far as I can tell but I'd love for someone to prove me wrong.

A Comprehensive Directory Server - The last suggestion I made was for Red Hat to somehow come up with an Active Directory killer. Work on FreeIPA has continued to progress and that is built on top of the 389 Directory Server which has also continued to progress. Both are sponsored by Red Hat.

I'm sure some will point out the upcoming Samba 4 release but that is just a clone of AD, not a directory server for managing Linux machines. I also have to question the validity of adopting Microsoft's technology design. Are we really considering taking a page out of Microsoft's play book to adopt and extend?

FreeIPA had a v2.0 Alpha release last month although I don't know anyone using it so I'm not sure how useful it is yet but I'm definitely glad to see they are working on it.

While there are a few other directory server products for Linux on the market, none of them seem to have gotten much attention nor many noticeable deployments. Microsoft really has been gaining speed with Active Directory and it seems that most all third-party software of any complexity supports it. Take Red Hat's own RHEV for example.

Perhaps it is time to invest more resources in FreeIPA.

New Suggestions

Virtualization - Red Hat has decided to become a full-blown virtualization company by merging with Qumranet, releasing KVM support in RHEL 5.4 and by releasing Red Hat Enterprise Virtualization for Servers last November. RHEV for Desktops is currently in beta and expected to be released sometime this year. Unfortunately the current release is based on a Microsoft Windows-based management app but they are in the process of changing that. I have taken RHEV for Servers for a test drive. It seems to be a fairly competent product even in its current release BUT it seems designed for higher end environments. RHEV has some maturing to do because even though it has the feature set, it can be quite cranky.

What I'd like to see is Red Hat release their RHEV-Hypervisor product freely both in binary and source form... which the later is already available I believe. They'd also have to tack some sort of minimal web-based management system on top of that. Basically what I'm asking for is a version of RHEV that is similar to VMware ESXi. I realize that Red Hat is almost there because much of the underlying support software is freely available (KVM, libvirt, virt-manager, virt-install, etc) in Fedora... but not really as a very light-weight turnkey solution. They will probably point to the fact that they sponsor oVirt and that it is out there but again, it is quite more complicated than a small to medium sized business needs.

While KVM has been adopted by most Linux distributions and is gaining mindshare quite nicely, I don't think that is going to do much for RHEV sales. I do not see Red Hat making a dent in the greater virtualization market except for businesses that are already mainly using RHEL. In order for Red Hat to have a chance at making a dent in the virtualization market, they are going to have to release an entry level product for free... that is useful enough to be adopted by small and medium business... and then RHEV will sell to those who want more features. VMware is the 800 pound gorilla and while they actually have some competition, the VMware product line is very mature, feature filled, and well liked in the IT industry. I don't care how well RHEV does on merit, it will not truly compete until it has some underlying marketshare built by an entry level product that has Red Hat's name on it. Maybe they can do it without that, but it'll take much longer.

What about containers? - I would also like to see Red Hat recognize OS Virtualization aka containers in some way. LXC is quickly developing in the mainline kernel and both Parallels/OpenVZ and Linux-VServer are out-of-tree patches that have been doing quite well for a number of years.. Being able to do both containers and fully virtualized machines sounds like a winning combination to me. Right now the only two products doing it that I'm aware of are Proxmox VE and OpenNode.

The Impending Canonical Threat - While it remains to be seen if Canonical will ever have a business plan that actually launches them into profitable territory, they seem to be doing quite well in the mindshare department. I do not care to go into a long analysis for why that is... but the fact is that Ubuntu's desktop popularity may creep over into the server space. The question is, should Red Hat be concerned? I'm not sure. To me, Ubuntu is not an enterprise distribution. It might be fine for small to medium sized businesses but I don't know if they can enter into the enterprise area and compete with Red Hat. Some will say they are already there and that yes, Canonical can compete with Red Hat at the Enterprise level. I'd definitely like to see Red Hat have more competition because that will improve their products.

What remains to be seen is if Red Hat will decide that they actually care about small to medium sized businesses and actually go after that market. They would also have to show they care about the desktop again and that they think Linux can be competitive in the desktop market. I'm not sure if Red Hat is ready to go there or not and yes that market space is definitely filled with some profit black holes. While I'd like to see Red Hat move into the lower end markets, I'd only like to see them do so if they have a reasonable business model that won't devour all of their profits. I'm not sure if the time has come yet but one has to ask, if not now, when?

With regards to Canonical, I'd rather see Red Hat publicly act in a slow methodical way rather than in a last minute reactionary way.

Conclusion

I don't want this article to sound overly negative. Red Hat has truly been and remains THE beacon of success in the Linux market. Most other Linux companies have been chasing Red Hat's tail lights and following every turn they make. Red Hat has definitely been showing strong leadership by sponsoring emerging technology both in the Fedora Project and elsewhere. They have kept the Free Software and Open Source faith. The good news is that they have remained successful doing it.

Good job Red Hat and Happy birthday RHEL 5!