Recently, Pluralsight published a blog Going open source at work: How to convince your boss which covers aspects that might prevent users from adopting open source and provides reasons on how to help, well, convince your boss. I couldn’t disagree more with the post. That’s not to say however that open source might not be a good fit for your organization, but the points given are extremely one sided…as I seem to typically find with most open source proponents.
The first point the author makes is that linux is being used more and more and even uses Facebook as an example of companies using open source to run their business. The biggest problem with this reference, and one so many business people bit on, is that your business is not Facebook, its not Google or Rackspace or any of the other companies making open source “sexy.” Yes, open source IS the right answer at companies like that, companies that have budgets to dedicate engineering teams (development, QA, release, dev-0ps, whatever you’d like to call it) to support operational maintenance for their systems. Said another way those departments are being paid to support a product that does not make the company any money. Now start-ups have successfully used open source software for years, companies much smaller than even a normal sized business, however keep in mind that start-up companies, in many cases are chalk full of software and QA engineers. These are the very type of people you need to successfully support such applications, but remember to allocate some of their time to support these applications, they can’t be working 100% of the time on new product features for your application and simultaneously support bugs or problems found in the open source application.
The next point the author makes is flexibility, and yes while the source code is available to you, the question you really need to ask yourself is does your company have the skill set AND time to devote to writing, updating, QA’ing and maintaining code to support these new features that you COULD build? If you don’t, then open source is not the right answer for you. The author uses Microsoft Office as an example. Can Office get pricey? Yes, does 98% of the workforce know how to use it and prefer it? That would also be a yes. Just because there is an alternative application, will YOUR business users be able to be productive with it? Is saving $200 worth lost productivity for your VP of Sales, Controller or receptionist? I’d likely take the bet that when you want something from people within your company, hearing they can’t do it because its a feature is not available or different than what they have been using for years, that it is not going to be an acceptable answer.
Control, which sounds like the author is really referring to licensing restrictions on commercial software, is also a bit of a hoax. How many different types of open source licenses do you think exists? 5? 10? 25? Try 43 (at least according to wikipedia). While many are corporate friendly if you are not redistributing the software, you should still have a legal department review the license agreement to ensure your company can’t be sued, even at a later date if the open source project is acquired by a commercial company who then changes the license. Are you ready to take on the cost of legal fees if you use external counsel or additional time worked for internal legal staff to review licenses and keep up with changes to licenses?
Speaking of taking on cost, cost of ownership may be the biggest myth in open source software. It is NOT cheaper to use open source software, if it is, its by a slim margin. Open source software is not a product, while some projects can be used to perform a specific tasks, it should generally be thought of more as a platform to build on. You also have to ask yourself does it integrate with your existing systems? What is the learning curve? Who is going to support bugs? There is a people cost to open source, and finding good people to support those applications is not easy. I often use the refernece of finding a good linux admin versus a good Windows admin. When it comes down to it, I can find a good Windows admin tomorrow (not to take away from their skills, there is just simply more of them out there). Having been part of searches for linux admins/engineers I can tell you that you are unlikely to find a good one quickly.
That brings up support, the authors last point. He states there are options for supporting open source such as writing to mailing lists and forums. With all due respect, those are not viable support options for enterprises who might be relying on this software. For fringe use, or low impact applications that might be viable. But if this application is going to support the entire organization, or even someone high up the food chain at your company do you really want to rely on a mailing list or forum for help when the CEO is standing over your shoulder? To be fair, many open source projects do offer commercial support, and there are companies whose sole business model is providing support for open source applications but your cost benefit argument quickly dissipates if you need to go this route. Beyond support, what happens when there is friction within that open source community? Look no further than CentOS and Nagios. It was not long ago that people were pushing the panic button over friction with CentOS and jumping to other distros (a slightly older article here, only 3 years old) or developers “leaving” an open source project and creating a new one like what happened with Nagios and Icinga. Is that acceptable to your support team and organization to just switch operating systems or applications because of what boils down to a lovers spat? Can you afford to lose support, or the amount of time and money previously invested in an open source project if its abandoned?
Another interesting support point, just because a application or company focuses on commercial software, let’s pick on Microsoft and VMware here, I would actually argue that the support provided by the community is far and away better than most open source projects. In fact, its the support from the VMware community that led me to recommend ESXi back in 2008 over Xen Server. When we ran into problems with ESXi there was, and still is a huge community offering support and suggestions. These commercial software companies also have the resources to dedicate to maintaining and active knowledge base and documentation, something I find that tends to be lacking in open source projects (dear open source developers, man pages are not documentation).
Along the lines of support and abandonment, life span is another tick in the pro column for open source. Are there older applications that work? Of course, but are they supported? There are just as many abandoned open source projects, and I’d be willing to be more than there are active ones. Windows XP commercial support has finally come to and end, the right move would be for Microsoft to opensource it but at the end of the day, the need for XP was really killed with Windows 7, there isn’t much reason not to upgrade to Windows 7 in 2014 if you aren’t there yet. Oh, what’s that? Your custom application your engineering team build won’t work on 7? They don’t have time to update it? In that case will they really have time to update it if there are changes to the open source project or operating system they built it around?
Summary (also in this case the tl;dr version)
Now that my anti-open source rant is over, lets be clear – I am NOT against open source. There is a time and a place for EVERY application, every operating system whether commercial or open source but it comes down to YOUR specific technical AND business requirements. Old school corporate IT departments are dinosaurs, they do need to be more nimble but you have to, again, look at your specific business and what your needs are. Can you support the application or operating system within your service level agreements? If you can’t can you bring in people who can? Can you afford that turn over in? Can it be adopted and used efficiently by departments outside of IT? Are you will to allow an appropriate amount of time in your engineering team to keep up on changes? Remember not everyone is a computer geek. Would it be nice to save money of Office, I’m sure every company would love that but does OpenOffice work with your existing applications? What would the cost be to replace those other applications it needs to integrate with? Most importantly, is your organization ready to invest in open source the way it needs to be?
Remember, open source is a platform, and more than likely the support burden will be placed on your organization. Always, always, always consider the business as well as technical needs.
The vCenter support assistant is a free appliance from VMware that assists in submitting support cases to VMware while collecting relevant logs. This should be installed in every environment now BEFORE you have a problem! The installation is very easy:
Indeni Dynamic Knowledgebase is a “game changing” network device monitoring tool that allows administrators to scan their network and monitor devices based on real world device profiles. In addition to monitoring your devices, Indeni also provides remediation information for each alert. There are two versions of the product today – the Dynamic Knowledge base (i.e. enterprise/full version) and the stand alone Indeni Health Check which is a free product that runs on a standard desktop OS with Java installed. What is interesting to me about Indeni is that they have taken the next step in vendor agnostic monitoring and built in solutions to errors, dare I say a “big data” like approach versus configuring your basic ping or service monitors and then leaving the troubleshooting up to the engineer/administrator. The features from Indeni include monitoring, device backups, profiling (think host profiles in VMware for validating configuration settings), change management and troubleshooting.
For the purposes of this post, I have downloaded the trial of Indeni Dynamic Knowledgebase which is delivered as an ISO image that includes a hardened linux OS. The user guide can be found here. Once downloaded, place the ISO in a location available for you to install, in my case I have uploaded to one of my datastores. I then created a VM and selected the ISO image for the CD device. Once booted, you go though a pretty general linux setup to set the IP information and then let the installer configure the necessary components. One item to keep an eye for is that you need to manually restart the system once the configuration is complete (seem’s like that should be automagic). Once the restart has finished, you can access the web gui at https://ipaddress:8181
From here, you can log in with the default Indeni admin/admin123! user. Accept the license agreement and decide weather you want to enable device deletion, for historical reporting purposes you may wish to select ‘Disallow Device Deletion’ or you may require this setting for other reasons; I chose disallow since it cannot be changed. Once logged in I added a switch to verify Indeni could communicate with my network devices and it was able to successfully log in and verify that my device was online, with no alerts. Once given a few minutes, it will start to generate recomendations for the device(s) it is monitoring. You can see here it suggests I create a dedicated user account for Indeni to use across all my monitored devices.
The one area I was most interested in testing was the device configuration, backup and change tracking component however I have been unable to get this to work. While my device appears in the monitored list, and I configured a backup job, there are no actual backups in the backup location (by default /var/indeni/backups/devicename) so I am not able to try this feature out – maybe a bug, maybe I am having trouble RTFM’ing but in my opinion this should be the easiest part to configure.
Indeni is certainly on its way to creating an interesting product offering for managing and maintaining your network equipment. The built in knowledge base, ability to monitor and to search config backups alone make this an interesting option for device management. Stay tuned for a review of the free desktop version of the product.
First, I want to apologize for only having 1 blog post in January. In an earlier blog post, I stated that I was kicking off 2013 with a trip to Seattle, WA. Seattle is home to the headquarters of Isilon, one of EMC’s more recent acquisitions. While the initial purchase and sale went through a while ago, it was only recently that the installation, sales contracts, and support system was converted over to what EMC has been using for years. For 2 and a half weeks, I went on site to help the Isilon Remote Support staff understand the inner workings of the new ticket system known as “Service Center” and how it related to their existing system based on SalesForce.
Decisions were made as to what would be converted over and what wouldn’t be. The existing SalesForce database would still be available to support, but in a read only instance. This would be because only cases that were still open on Saturday January 12th were going to be converted over. Older closed cases would still remain in the old system for reference, and new cases would be created in EMC’s system only.
I’ve been in support long enough to know that not every migration goes according to plan. As far as this one went, there weren’t too many major issues. The minor ones were remedied quickly and it seems like everything is finished. Existing EMC customers should be familiar with the online support interface. Your Isilon arrays would show up under existing sites if the address on file matched, otherwise you may need to search for new site IDs to follow.
The culture over at Isilon is still very much like a startup and I loved it. Moral is high and it seams people really enjoy their job and care about the customers they help. It’s no surprise that it was voted one of the best places to work in Seattle. With any major change to a business process, it is going to take some time for those involved to get used to it. It seems like everyone was able to adjust quickly and by the end of my stay, the support engineers seemed well adjusted to the new system and processes.
All in all I think the conversion was a huge success and I hope you will join me in welcoming the Isilon Remote Support team to EMC.
On a side note: I had some free time to see the sights of Seattle. I posted a large collection of photos from my trip (broken down by day) and I hope you enjoy them.
Source: Thulin Around