Kung Fu Apps

January 9, 2007

Just yesterday I commented on what I said was an “almost irrelevant” deal by Intel that bundles semi-related software —– such deals really serve the purpose of an attempt at furthering product awareness and easy product acquistion rather than serving as real solutions that address business issues.  Large numbers of business customers weren’t screaming for an easy bundle of traditional web 2.0 style applications like blogs and wikis.  Tracking RSS feed subscription statistics for the public Internet is one thing, but tracking those kinds of statistics internally for enterprises might be a little too early.  Continued proliferation of RSS throughout enterprises will certainly change that.  Regarding SuiteTwo, I went on to say:

“Simply creating a bundle of un-targeted, but related software as valuable as each application may be individually, usually doesn’t create any significant total value in sum.”

One of the companies in Intel’s SuiteTwo deal is Newsgator, a company that counts me as a happy paying customer incidentally.  Newsgator just announced a deal today with Motorola’s Good Technology Group to deliver content via RSS on handheld devices.  This is a significant deal with a clear and growing number of targets in mind.  Extending RSS support to handheld devices is an obvious and natural extension to the communication and content paradigm.  (Wow, that sounded like it came from a bonafide shill.  I’ll stop I promise.)

Developers of traditional enterprise software are continuing to output RSS for internal consumption and more consumers of RSS will gradually adopt these new delivery mechanisms.  But I continue to believe that the killer app in regards to enterprise RSS is not limited to people using Technorati for persistent searches of their company/competitors or for accessing public Internet or semi-private content —–  the killer app with RSS involves enterprise mashups.  More specifically, when developers of enterprise software begin to output content using generally accepted ontologies that consitently describe data, the world will be a happier place.  In some respects, FOAF-like semantics need to evolve within the business world in order to coordinate “machine agents” as their so often called.  But that level of coordination won’t happen tomorrow and agreement is a hard thing to come by.  So in the mean time, an enterprising startup will come along to interpret the syndicated output (and possibly the APIs too) of various vendors to provide more meaning and value based on this new form of content consumption. 

January 7, 2007

Intel announced a “Web 2.0 Business Internet Suite” in November 2006 called SuiteTwo consisting of a bundle of semi-related applications from a handful of vendors.  The applications include MovableType blogging software from SixApart, wikis from SocialText, RSS feed reader support from NewsGator, and RSS feed analytics from SimpleFeed.  SpikeSource, who we have written about before, is delivering the suite through their testing and support mechanisms.

I’m a big fan of bundled software —— where particular solutions are in mind.  I spent the majority of my adult life working large OEM deals for/with big software companies.  I know with certainty that none of the deals would have produced material amounts of revenue had particular solutions not been targeted.  In it’s current form, SuiteTwo is a solution with no target.  Simply creating a bundle of un-targeted, but related software as valuable as each application may be individually, usually doesn’t create any significant total value in sum.  Sure, ensuring compatibility between applications is important, but collaboration alone doesn’t drive business or solve problems.  Applications must be integrated into core business applications to have any significant relevance.  Missing from this suite of products is something from a company like OpenSpan that can tie line-of-business apps into collaborative applications like SociaText and MovableType using high-level integration techniques.  Otherwise, applications become islands of information with no connection to business applications.  With no little or no connection to the line-of-business, applications fade to grey very quickly.

Show me a “Web 2.0 Business Internet Suite” that addresses specific issues in mortgage loan processing, accounts payable/receivable, and human resources just to name a few.  Most CIOs will ask one question:  what problem does this solve?  And the answer can’t be something along the lines of, “Your employees will be able to collaborate better.”  Show them how lengthy receivables cycles can be reduced or payables can be made faster thereby earning early pay discounts — that’s relevance.  Take one look at the SuiteTwo site about “Why Enterprise 2.0?”  There’s not a single measurable bullet point amongst the list of reasons to buy their software.  This is my favorite one regarding employees:  “Participation leads to improved morale.”  Right.  Intangible bundles are usually the result of Barney relationships:  “I love you and you love me,”  but how do we together make some money?

January 4, 2007

When EMC acquired RSA in June for $2.1 billion and IBM acquired ISS in August 2006 for $1.3 billion, I jokingly said that the Gartner and Meta/Giga/Alpha/Beta types would say that the “security market was heating up.”  There have been a couple of other smaller acquisitions since that time in the space, but Cisco confirmed yesterday that the the spam and anti-virus aspect of the security wheel was a long-term investment considering the $830 million acquisition of IronPort. 

IronPort had raised $94 million in venture capital from a gaggle of investors and had only booked an estimated $25 million in revenue during 2005.  So a price tag of almost $900 million constitutes a jeroboam of a deal.  And it makes it clear that Cisco is placing a big bet on a huge problem.

Hat tip:  The Alarm Clock

January 3, 2007

Collaboration is most certainly a key part of the new enterprise whether it’s in wiki, blog, or traditional application form.  CorasWorks, based in Reston, VA, recently raised $4 million from Edison Venture Fund as was reported in the Washington Business Journal for their impressive suite of applications.

CorasWorks has essentially built a series of packaged applications on top of Microsoft Sharepoint.  I’m bullish on the notion of collaborative tools that offer easy-to-deploy modules with minimal setup.  The difference between solutions from companies like CorasWorks and the other new-age collaboration vendors is major:  CorasWorks develops specific business solutions with their applications.  The list of solutions in their WorkPlace suite is impressive and it’s not just limited to collboration centric solutions.

Most of the new-age collaboration vendors fail to get this key element in my opinion —– businesses buy solutions that affect their business needs.  Collaboration alone is not a sufficient value proposition for customers —– it has to be integrated with their core business processes to add value. 

The company claims 700 customers constituting 500,000 deployed seats.  That’s impressive for a 3 year old company.

December 21, 2006

Seagate Technology will announce today that they are acquiring Evault for $185 million.  Evault provides hosted, online backup solutions for small-to-medium businesses that competes with the likes of LiveVault who was acquired by Iron Mountain in December of 2005 for $50 million.  Evault will be bringing 8,500 clients with them to Seagate according to Forbes.

Kidaro, an Israeli-based desktop virtualization developer, recently announced that the company secured an additional $10 million in venture capital for their foray into new ways of delivering applications.  Opus Capital joined Storm Ventures and Genesis Partners, existing investors in the company, in this latest round.

The company is doing something simliar to moka5, who we wrote about recently, in that they’re building on top of existing desktop virtualization platforms rather than trying to compete with the likes of VMWare and Microsoft. And distribution of virtual machines is the name of the game for both companies, but that’s where the similarities appear to end.

With Kidaro, IT Administrators provision workspaces for end-users that are delivered through the Kidaro client. Kidaro has a unique mechanism to deliver and present applications hosted in the virtual machines to the desktops that access them.  How they get the data to the clients is brilliant and borderline, well, magical.  Here’s an example:  the virtual machine being targeted for delivery and the desktop receiving the virtual machine both run Windows XP.  So why deliver duplicate portions of the operating system when it already exists at the client?  It never has made sense because it’s just a waste of network resources getting the data down to the client when most of the binaries may already be present.  But precisely how they’re managing to pull this off without requiring extensive computing cycles at the client is the significant part.

There’s another cool aspect to the product:  applications delivered in the Kidaro virtual environment appear as if they are running locally to the user.  But the working data itself is safely guarded, either because it’s at a centrally managed data center or encrypted when stored locally with the virtual machine.  So remote users who lose laptops with sensitive data on the disk won’t lose their jobs and won’t create a huge security risk for the company.  I just wrote about one solution to this problem yesterday while discussing Kazeon’s latest offering.  Mobile users are an inherent security risk and Kidaro’s solution is to not deliver the data to the remote machine in the first place or encrypt it if required. 

I’m very intrigued by this sort of innovation being spearheaded by Kidaro.  They’re leaping forward where companies like Citrix and VMWare left off.  Things are different these days compared to when Citrix had it’s moment to shine —– every aspect of hardware and bandwidth is much cheaper.  Kidaro’s application would have been somewhat absurd a short time ago given the change in hardware costs.  And at the end of the day, it’s inexpensive computing resources that enables newer ways of interacting with people through applications and data.  Anyone who thinks that the next generation of enterprise software applications, frequently dubbed Enterprise 2.0,  are solely based upon blogs, wikis, and other forms of social media are misguided.  Take one look at Kidaro.

December 20, 2006

Kazeon, a vendor who was traditionally focused on the storage management aspect of data classification, announced recently that they were adding security to the context of information management.  They announced a product offering and service, that appears to be an extension of existing technology, that addresses the problem of sensitive data residing on the edge of networks such as on laptop computers. 

This is a huge problem because laptops are getting stolen and discs are getting lost.  I have written about some of the inherent challenges related to information management in a portable world before and the short summation is this:  there’s no panacea for the problem because it’s a multi-headed beast.  Many losses of sensitive data are attributable to removable media, so there’s no application that can magically scan a CD-ROM full of sensitive data in my laptop bag.  But it makes sense for a company in this space to step into the security wheel and manage what is possible.  They’re already touching some of the data — why not extend the reach and then make sense of it from the perspective of possible security leaks? 

From the press release, it sounds like Kazeon is employing their crawler to search nodes at the edge, but it’s very difficult to maintain state and relevance in an application like this when there’s no code running at the client to keep up with the possibility of frequent changes.  An application like TrustedEdge does just that by deploying client-side code, but at the expense of deploying and maintaining client-side code.  There’s no easy solution for such a problem.

December 19, 2006

There was quite a bit of discussion recently about whether or not the traditional venture capital model is broken. Much of the discussion stems from the fact that Sevin Rosen chose not to raise their tenth fund given the less-than-optimal return and exit conditions. Some said that Sevin Rosen pulled a snow job on the NY Times to spin their story optimally because the facts of the story were something else entirely. Whatever the reason, it’s obvious that venture models are evolving. I would suggest that traditional VC models are changing in the same way that enterprise software companies and customers of such are changing. Venture capital models and enterprise software are evolving towards greater efficiency to remain competitive. Here are some reasons why traditional enterprise software vendors are in trouble in 2007:

1. Open Source Is Everywhere – Customers are getting acclimated, at varying levels, to using open source software. Proprietary models of 1990’s style innovation are melting like the glaciers. Using open source software gives an advantage to startups in terms of capital efficiency, but it’s not easy for a traditional enterprise software firm to shift to such practices. That doesn’t mean that they have to release the source code to their product, but shifting focus to open source could practically be a re-write. But the competition is out there and the pressure is on. Funding of open source software companies is up 131% in 2006 to 404.5M. MySQL is in, Oracle is out. (Microsoft SQL Server was never in by the way)

2. Installing Software Sucks – The fact remains that one of the primary reasons that traditional enterprise software is in trouble is rather basic: having to install software, whether on the client or the server, sucks. It sucks in the sense that installing software for purposes of self-hosted operation requires that it has to be maintained, monitored, upgraded, etc. That requires infrastructure and staff. Appliance based delivery addresses some of these concerns, but thin client, remote hosted, delivery of most software services is much more efficient on multiple levels. Service oriented models are in, servers are out. But if you insist on selling deployable software, offer an appliance-based delivery, purty please.

3. Distribution Sucks Too – For the same reasons that installing software is a chore, setting up and maintaining traditional distribution channels is expensive compared to the new world of try-before-you-buy and service-oriented delivery. Large, self-hosted applications need to be sold somehow and a salesforce can be expensive. Traditional enterprise products don’t get sold and installed magically. A large channel looks great on paper, but the majority of revenue likely comes from a small percentage of downstream partners. Salesforce.com is certainly innovating in the service-oriented model with their AppStore distribution model. It’s an on-demand application marketplace like iTunes is a marketplace for music. Brilliant.

4. Customers are Buying Complete Solutions – For those customers that choose to purchase the larger scale, self-hosted systems in house – they’re typically buying solutions from a single, larger vendor rather than from startups or midsize software firms. A small firm that provides a narrow piece of a larger solution must have a multi-pronged approach to solution selling. It’s difficult to sell one piece of the pie.

December 13, 2006

Yahoo announced on their blog today the release of a limited functionality enterprise search application available for download free of charge.  Such a freely available search application is clearly aimed at users considering one of the Google search appliances, but the actual effect will reach much further than Yahoo’s nemesis.  Enterprise search is a very crowded space.  It’s not just that every large, enterprise software vendor has their own search offering — it’s that several dozen startups in 3–4 different primary areas have their own take on enterprise search as well.  Some startups are focusing on search from the storage infrastructure perspective a la Kazeon, Scentric, Njini, and StoredIQ whereby archiving and data movement is a core part of the offering.  Other more established vendors that do more than just filesystem and HTTP search include Autonomy and Endeca — they do other useful content centric things with the data they find compared to the infrastucture plays.  And some vendors have a firm footing in the OEM search business like FAST and ISYS.

The real significance of this limited functionality release from Yahoo is that 1) It’s free.  2)  It’s a free product from brand name vendors.  3)  It’s a free product from brand name vendors that covers most of the needs of small businesses.  So any vendor that is selling their core product, regardless of particular niche, based upon search and discovery business drivers to smaller size businesses, will likely have to respond to the fact that a free product is now available from companies who know a thing or two about search.

The crawler portion of the Yahoo/IBM branded version of the application supports not only filesystems as source, but also allows HTTP urls to be added that can be included in search results.  The OmniFind Enterprise Edition supports larger business, enterprise class sources for the crawler such as FileNet, Documentum, DB2 Content Manager, Hummingbird, Sharepoint, and OpenText amongst others.  The concept of crawling an ECM repository and providing federated results is not new by any means — every major ECM vendor makes a point to be able to crawl their competitors respositories.  The Enterprise edition requires the WebSphere Information Integrator Content Edition (acquired from Venetica) that provides federated search functionality from the various ECM sources.

The Yahoo-IBM-product-name-is-too-long offering is not based on the piece offered as Yahoo’s Desktop Search.  The offering appears to be written in Java and uses Lucene for search and it’s offered on Linux and Windows platforms.  It also uses an SDK from Stellent (now owned by Oracle who has a very similar product) for processing the various filetypes.

I installed the application on Windows in a matter of minutes.  Like any other application that fetches large amounts of data, be prepared to wait for it to index the data.  For 1.5GB of PDF files, it created an equivalently sized index file although the application logged several messages about size limitations on two files that were 65MB+.  The search and results pages can be customized to a certain look and feel, possibly to mimic an existing intranet site to provide a seamless experience.  Yahoo Internet searches can be included as part of the query results as well.  There’s even an API, limited as it may be, to extend some of the functionality to custom applications. 

The free download of IBM Omnifind Yahoo! Edition (a sincere mouthful, really) can be found here.

December 11, 2006

I love turning that which is old into something new and useful. Opportunities for such occurrences are plentiful these days considering the rate at which hardware becomes old. Anyone who reads this blog with any regularity knows that I’m a huge proponent of virtualization software.  The ability to migrate virtual machines between physical hardware is so trivial that it’s almost mindless.  Relocating nodes of virtual machines would probably be finished copying from source to destination before you came back with a cup of coffee.  Managing virtual machines in much easier in general because of this level of flexibility.

Machine virtualization and appliance based distributions for software releases will become as ubiquitous as the silicon itself.  I think that’s a future indicator of another possible trend that I have observed with some clients already in the small-medium business sector:  rolling-your-own appliances using open source software.  I don’t know of a single, centralized place that describes some of these DIY business solutions, so maybe I’ll start adding them here as I find them for business use. 

Many small businesses already use open source solutions for things like firewalls.  The open source firewalls are all built on Linux or BSD, so the underlying operating systems are usually weathered and proven.  I use one personally called m0n0wall.  It’s one of the most stable pieces of equipment used for home and my business. The uptime is only affected by a severe power outage —– otherwise it’s up as long as any of the much more expensive Cisco equipment that runs the network.

A bright fellow by the name of Olivier Cochard adapted the code from m0n0wall and created a NAS appliance product out of it called FreeNas.  I recently bought a 1 terabyte internal IDE drive and threw it into an old PC to run FreeNas as a backup device for some of our machines.  So my total cost out of pocket to run FreeNAS with 1TB of data was $250.  10 years ago, there was no possible way of getting that much storage for anything less than several hundred thousand dollars. 

FreeNas .68 is the most recent release.  It’s still not exactly easy to install for a completely novice user and it still lacks a few features that are found in commercial applications, but over time this will surely evolve. Performance is quite respectable as well. I didn’t break out ethereal, but just eye-ballin’ it shows that it cranks along with massive file transfers.  Once the product does get installed, the feature list is hardly scant though.  It supports many features that smaller businesses need in terms of data storage.  If one’s clients don’t need all the higher end features that do things like snapshots and cloning, then FreeNAS is likely a very good candidate.

Download FreeNAS here.  Turn an old PC into something usable!