I really expected the HTC Dream (T-Mobile G1) to seriously dent the mobile device world. I had been using the Android SDK for a bit, but didn’t have hardware to test on, so I bought one. Great keyboard. Decent UI. Tiny screen. Slow processor. Lousy device support. Horrible network.

A little later, I really expected the Palm Pre to be the nirvana of compact, hyper-connected mobile devices. It wasn’t pretentious. It had a f%$#!%& keyboard. It was ahead of the curve in a number of features. But it wasn’t substantially different. Palm made the concious decision to keep the interface very similar to the old Palm OS, and let’s face it – user interfaces have evolved since then. But I got one. I like it. It’s a nice toy. If you want a compact, full-features smart phone, it’s still your best bet.

Motorola unveiled the Sholes over the summer. Beefy processor. Best-of-breed screen. Ridiculous connectivity solutions. A f%$#!%& keyboard. And CDMA (like the pre) so I don’t have to use the Ancient Telegram & Trash network. 5MP camera w/ LED flash. Ran Android (1.6 at the time). Super cool. At that point, they were still shopping for a vendor. I was cautiously optimistic it wouldn’t be a metro-only network like T-Mobile or Sprint.

Then came the onslaught of Droid Does during the baseball post-season. Speculation ran wild as to which phone it was, beneath the hype. BGR scooped it and pointed it out as a Sholes running Android 2.0. Thursday last I received mine.

Metal. Everywhere.

Ridiculously clear screen with outstanding pixel density.

Incredibly fast processor.

A f%$#!%& keyboard.

Seamless integration with all of the stuff I use (e-mail, calendar, contacts, etc. etc.).

Transparent movement between WIFI and the Verizon 3G network.

Incredibly fast processor.

Deep interface built atop a fully-accessible Linux system.

Scads of customizability.

Surprisingly good camera with shockingly bright flash.

Oh, and an incredibly fast processor.

If the pricetag is $100 too high for you, Verizon is also offering an HTC-based version called the Eris with no keyboard, more plastic, and a mid-level processor, with the same interface and general feature-set.

Oh yeah, and when you send e-mail, it’s not tagged “Sent from my <BlackBerry|iPhone|Other Pretentious Device>”.

But, since most people seem to enjoy those things: This post authored using a WordPress app from my Droid.

[UPDATE: Avaya did indeed win the auction, for nearly twice their stalking-horse price. We'll see how that plays out.]

This Friday, Nortel’s enterprise products division will be auctioned. Avaya is the stalking-horse, with numerous other companies additionally submitting bids. If history holds, Nortel Enterprise will not be Avaya’s, but belong to someone else (despite all the buzz Avaya and its partners are making). Regardless, that someone will acquire the best enterprising data switching platforms made.

Hands-down, the ERS 5000 -series has continued to be the best integrated UTP switch, since coming onto the scene many years ago. Beefy backplane, generous uplinks, flexible stacking, decent price-point. I’ve baked switches from every vendor willing to send me gear, and no one can handle the battery like an ERS 5000.

The ERS 8600 chassis-based platforms also make a mockery of the competition. Robust, redundant backplanes; extremely flexible slotting options; ridiculous feature set; decent price-point. Better marketing could have really pumped this line, as a number of the slot options were best-of-breed. Again, I’ve baked a number of chassis, and the 8600-series delights on most all of the benchmarks that matter.

There are lots of other diamonds in the enterprise space. Quite a lot. I’m not a “loyal” customer of anyone: I believe in market Darwinism. Unfortunately in this case, the best products were latched onto a fiscally irresponsible company, led by brain-dead opportunists who from the outset have undermined the legacy that is … was … Nortel.

I look forward to Friday – to see who’s getting the chance to own the best-of-breed enterprise networking products, and hope when they sit down and take a good look, they realize what an amazing set of assets they have, and Do The Right Thing with them.

Regardless, Nortel is a less-than-skeletal image of its former self, and a venerable multi-century legacy is all but a footnote.

This digital picture frame is the best I’ve ever seen (You can find it MUCH cheaper other places).

It has a USB-computer interface for uploading into its built-in memory, takes a large swath of media cards, and even allows you to plug in USB thumb drives (or external drive caddies)! It also supports MPEG, QuickTime and AVI movies as well as MP3 audio files you can use to soundtrack your slideshow. The color is crisp, the contrast ratio nice, and it even handles my plethora of panoramas perfectly.

I’m not a fan of digiframes in general, because they don’t do everything they should, and are usually horribly overpriced… This one does and can be found pretty cheap.

RedHat released a version of their RHN product (dubbed SpaceWalk) open-source. Fedora, of course, quickly had their own site up about it.

I’m glad Levanta already died, because this would’ve destroyed them. RedHat is wise in its timing, as I know Novell is gearing up to really take on this market segment, as was evident by their recent acquisition of PlateSpin.

Nessus is the security scanner. That’s not just a tagline, it’s the truth. Yes, other people make scanners. Yes, other companies make tons of money off of their scanners, but having used ALL of them (yes, every single purported ’security scanner’, that I have ever heard of, for the last 10 years or so. ALL), Nessus is the “best”.

Historically, one of its major values was its cost (free) as well as its source license (Free/Open). Cost is still, generally, free (restrictions and various non-free necessities apply) … but the source, and the product, are… not. This may upset some to the point of refusing to use it. I’m not one of them.

Nessus is very light-weight, its rule language (NASL) is very intuitive and powerful, and the sheer volume of support and flexibility provided is exceptional. It finds things other scanners only dream to. It can (not by default) work very stealthily, not setting off some IDSes. It can be configured not to destroy the systems it’s scanning. No one builds a better scanner. Truthfully, I wish someone would: Not because I don’t like or want to use Nessus, but because the ecosystem is very homogenous in this space: You either use Nessus, or you may as well find a dowser with a divining rod to point out your vulnerabilities.

Over the years, one of the things that has been failing with Nessus is the interface. I enjoy command-line interfaces as much as the next UNIX junkie, but when you need a tool you can put in the hands of Joe Average, it has to have a graphical interface… It just must. Historically, back when Nessus was free and Free, there was no shortage of Tk/GTK/Web/Qt/etc. etc. interfaces: A lot of those are still around, but are generally handicapped from the newer features. When your license prohibits reverse engineering, it makes it hard for people to maintain their interfaces to your product.

Recently, I re-acquired my main scanning system (after a stealth project became production, literally overnight (over a year ago)), and decided to upgrade everything… Including Nessus. The upgrade was flawless. It found my old install, upgraded it, re-registered my feed, etc. etc. Very slick. I then upgraded the client on my laptop. That was not so smart of me. Following a common trend, Tenable has outed the old interface in place of a “new” Qt-based interface. I have nothing against “new” or “Qt”, but the new interface is missing dozens of previously-exposed features and one… one very IMPORTANT feature has been removed from both the Qt version and the CLI version: html_graph. It appears there is no html_graph in Nessus 3.

What is html_graph? It was a report export format that presented an HTML report of the scanning results along with pie and bar graphs giving you a visual representation of what the report contained. Very very very important when presenting this information to people wearing ties.

I searched the Interlink: No vast outcry! Just one, lonely, unanswered post, reporting the problem and asking for help. I would guess that because he wrote in all lower-case, no one wanted to tell him the harsh truth: You can’t with Nessus 3. Nope. No more XML output. No more HTML+Graphs. No more. I’ll hypothesize this is because of the license changes: They can’t link against a whole slew of Free libraries that do this work for you, because of the shift from libre to locked. That’s just a guess. At first I thought it was just in the Qt-interface, which could have been simply a CBA or porting issue, still on the TODO list: But when I saw the neutered CLI – without XML or html_graph – well, that’s indicative of politics as opposed to growing pains.

So I don’t know, as I’ve said, why these are gone. But they are. Thankfully, as of right now, I can still convert the Nessus 3-generated scan files (NBE) into html_graph and XML using an older version of the Nessus CLI (version 2.something), but it’s a pain, it’s an extra-step (okay, okay, I automated it with 2 lines of Perl, but it’s a pain to someone!!), and it devalues an otherwise very very valuable tool… For if you cannot show pie and bar-graphs to those wearing ties, on-demand, the fact that you have found 4781 security problems is moot.

[NOTE: This was written a while ago]

[UPDATE: Levanta is toast].

Remember LinuxCare? That ambitious start-up back-in-the-day that wanted to revolutionize Linux support? Yeah, they died. But from the ashes rose Levanta. Unlike LinuxCare, Levanta has been studiously working on making proprietary products to help Linux environments self-support.

While unable to bake one personally (they’d love to ship one with a PO in hand and then take it back if I don’t like it), I did get a dog-and-pony show with a sales rep and an engineer that was quite impressive. Complete birth-to-death lifecycle management of server deployments (metal or virtual) using PXE and installing OS’s on systems automagically, managing update deployments, etc. etc. etc. Basically anything you can do with RPM (updates, installs, rollbacks, etc. etc.) and TripWire (change detection/management) in a box, on a larger-scale, with a nice GUI.

The product was impressive, but as usual, I wasn’t impressed with the required coinage.

Pretty quickly I developed some proof-of-concept soft systems that provided 75% of the non-GUI functionality using true NetBoots instead of the PXE-install-based methods the Intrepid uses. I also found a lot of other tools available including Cobbler that could help make a low-cost, high-value system of similar stature fairly easily. When thinking about virtualization, other tools such as oVirt, Virt-Manager, and Con-Virt could certainly be leveraged to provide and in-house solution to the TLM dilemma.

If you don’t have the ambition or skills to glue together your own, the Levanta Intrepid is by far a best-of-breed appliance for total-lifecycle-management of your RPM-based Linux server deployments.

[NOTE: This was written some time ago]

There are a lot of application switches out there. Large datacenters and server deployments can’t get away from them, but the small and gawky -sized deployments simply can’t afford them.

The Foundry ServerIron 4G is trying to change that. Priced a little out-of-range of small datacenters but still affordable if essential, the 4G offers nearly all of the features of large, enterprise-grade application switches in a 1U form-factor.

I didn’t get my hands on a -SSL version,  but did bake a non-SSL 4G for a couple weeks and it really performed well. Tossed a number of different scenarios and configurations at it, and everything worked great. One of my favorite features is the ability to NAT the requests/responses to/from your application servers, allowing them to act oblivious to the fact there is even an app switch there at all, even if they do tricksy URL rewriting (like the bane of rewriting, BlackBoard the Behemoth).

I’ve used quite a few devices in the application switch space, and the ServerIron 4G is really the only thing in this space that’s worth its price for small/gawky datacenters and server deployments.