by Jim Romeo

CTO Strategies: Greg Royal of Cistera Networks

News
May 28, 20088 mins

Cistera Networks helps tie business applications and phone systems together. Company CTO Greg Royal explains how the company looks for scalability, integration, and agility, through measures such as minimizing the desktop software load, and Which open source software he finds useful.

Greg Royal founded Cistera Networks six years ago to provide a convergence server platform that connects a company’s enterprise applications to telephone users, using voice over IP. Greg continues at Cistera as CTO and EVP, and writes a blog We spoke to Greg to discuss his company and its embrace of Linux in his industry.

Can you tell us a bit about your business and how you embrace Linux?

Cistera Networks is a software development company based in Dallas. We build enterprise application platforms for IP Communications, specifically Cisco Call Manager and Sylantro Hosted Platforms. As with most enterprise IP PBX platforms, we make extensive use of Linux in our products. It is an extremely robust and scalable platform to deliver an application appliance strategy.

Linux forms the basis of our enterprise IT strategy along with a comprehensive Cisco network infrastructure. Linux came about five years ago as a necessity, and has subsequently turned into a strategy for us.

In the early days, Red Hat Linux was used to provide Web site capabilities, as well as email, file and print services for a small number of client machines The Internet provided us with a remarkable amount of support for the various challenges we had in building out a cost-effective infrastructure as well as with the flexibility to try different service offerings and capabilities. These included CMS systems from  Joomla, Zope, and a trouble ticket system, RT. Other systems included CVS code repositories and Antfarm build management.

As our company grew larger, it was necessary that we move toward more robust commercial offerings However, we did not want to lose the flexibility that moving to a closed proprietary system would require.

For example, a key requirement was that we remain a multi-client environment, which included Mac OS X, Windows XP and Red Hat Desktop because of the various requirements of each department. The hard-and-fast rule was unless there was a very good reason, an application must run in a cross platform browser like Firefox. For this reason, we moved to Zimbra Enterprise for Groupware, Salesforce.com for CRM, and Joomla for all internal content management. The next major shift planned is the move to Oracle Financials on Red Hat Advanced Server.

How has Linux diffused into your enterprise and who do you monitor and manage your Linux boxes?

Linux touches every part of the organization The core enterprise systems number 12 servers running Red Hat Advanced Server. The IP Telephony system is Cisco Call Manager but also includes three Trixbox Servers for various services. Zimbra Enterprise on Red Hat Advanced Server provides Groupware.

User Management and Access management is provided via a Radius server which also contains the OpenLDAP server. Monitoring of the servers is done by Nagios, but we also have a VoIP QoS System from ManageEngine for our SIP servers.

Because certain groups look after their own departmental requirements (e.g. R&D owns code management and documentation), the framework of standardization is small and flexible. For example, we require LDAP support and it runs in Firefox Otherwise, all other management tools integrate into the Linux platform. This significantly reduces the overhead and cost of supporting various applications.

From your perspective, what has open source meant to your business of managing your IT enterprise?

Linux for us is a core IT strategy – rather than simply an adjunct to a Windows network – because we don’t want to live within the limitations of a Windows-centric network. Quite simply, using Linux means we can do more for less. We maintain two database standards (MySQL and Ingres), which means we do not apportion a large part of our budget to OS and DB licensing, but rather toward getting various projects and capabilities up and running.

From where you sit, where are proprietary solutions headed? Is there plenty of room for open source and proprietary solutions to coexist peacefully?

There is definitely room for both proprietary and open source platforms to coexist. While the quality of the offerings available for Linux has increased dramatically in recent years and is now supported by all major software vendors with the exception of Microsoft, there will always be companies that are comfortable with Windows-based solutions – just as there will be companies comfortable with Unix systems However, it is becoming harder to justify the cost of implementing and running a proprietary solution when there is such a high quality available at a lower cost. We also question the wisdom of tying specific applications to desktops. What we may lose in functionality on the desktop – but AJAX seems to remove that gap too, we more than make up for in the reduced maintenance cost of the desktop Our single desktop image for Dell laptops now consists of Firefox, Microsoft Office and Cisco VPN.

Let’s talk specifically about Total Cost of Ownership (TCO). How is TCO changing for those who go Linux, that others in the IT community might not realize or deny?

The biggest observation that I can make from hiring engineers is that Linux-experienced people are typically of a higher skill set. I look for people who run Linux servers at home and are not afraid to try things and test things at home. We run a large test environment made up of different configurations We don’t worry about the licensing cost and hardware is very cheap these days to run big test systems.

Engineers are encouraged to try out new software, especially open source to see what is possible In fact, we pilot open source projects to see what we actually need, and from there we may go and purchase a commercial implementation. The key is that we have a much better understanding of our needs from these trials.

Open source means we can try things without worrying about the cost and it means a better, informed decision when we do decide to purchase.

As the Linux application environment improves dramatically and the number of service offerings increase, our ROI also increases. We don’t need to make decisions quickly and hope they are the right ones. After all, a 30-day trial can only get you so far.

We are able to get onto new projects faster and keep the interest of the engineers.

Finally, a Cisco/Linux combination is very compelling.

How do you deal with different client requirements in a not-yet-Linux world?

There are a number of desktop-bound applications that require specific platforms. Fortunately, these are becoming rarer each day. Applications like Visio as well as certain network management tools require we have a population of Windows XP laptops and servers For example, the physical security and access control is Windows. We basically require that the dependencies on desktop OS be kept to an absolute minimum.

Getting out your crystal ball, for what types of organizations and enterprises is Linux going to be making the biggest impact in terms of cost and operability?

The companies that will most likely benefit from a Linux strategy are primarily young companies that don’t have legacy infrastructure to move forward.

Many startups obviously are finding benefits from starting out with Linux in the core and they are increasing. Technology companies are also prime candidates to embrace a Linux strategy If they have a strong Solaris or AIX background, then there is already a resource base to tap There is no innovation currently occurring on the desktop, perhaps with the exception of Mac OS X, all of the new tools are AJAX or Ruby-based; therefore, there is no preponderance for Windows desktops Companies that want to embrace a Web 2.0 future can use Linux as the core enterprise platform.

What are some things that you see from the seat of a CIO with respect to Linux, that not every IT person sees?

I am constantly asking the IT staff to think about the cost of desktop dependent functionality. It seems that any time you create dependencies on the desktop, you drive up the cost of managing infrastructure By deploying and managing a Web interface you only need to deal with the issues at the server level, rather than managing the desktop image of hundreds of desktops – a function that we did frequently at previous jobs. This also applies to server-to-server functionality. We constantly look for SOAP – or SOA-based technology – to ease the integration requirements. Removing dependencies between databases and applications is a big priority.