Everyone’s talking about open source and open standards now-a-days but hardly have I come across any who knows what exactly it contains and what the importance is. I’m putting this post in my blog today as I desire for people to have a better understanding of what is the difference between these two terms. Let’s get a first hand understanding of what these terms mean.

Open Source is a methodology based on which the design, development and the distribution of the developed software is carried out along with its source code. On the other hand, Open Standards means a royalty-free distribution of technology which has been formalized and approved by committees that are open to participation by all interested parties. For example, W3C is a body working on standards for Web Technologies. Click here for more information.

The word “Standards” means a set of guidelines to which a lot of people have agreed upon. Putting this definition in the context of software, “standards” allow a company to pick and choose from competing vendors and interoperate their systems without being pinned down to one of them.

Many think that open source software offers the same advantage, but that is not true. Open source means that you have access to source code for inspection and modifications but it may fail interoperability. This does not mean that open source is not good. There is lots of software available which are better or provide a similar kind of functionality that paid software will do. And the best part of open source is that it is free and comes with the entire source code. But what happens when you need assistance? Is the assistance free or is there a price tag? We all know that there is lot of stake involved in providing support. If you are thinking that support means training people on how to use the product or software, you are mistaken. There is more to support then that. It involves product trainings, more research to fix bugs and vulnerabilities, distributing patches and updates, etc. There may be companies that provide support on open source platforms and have no tie-up or relation with the parent company that designed and distributed the software for free. They may be able to fix small issues since there is access to the source code, but imagine a situation where there is something wrong at the core level and requires dependency on the original developer. Further to this, there is plenty of evidence that suggests that open source does not necessarily equate to cost savings. There are still licensing fees, service fees, implementation fees and support fees associated with it.

I would feel comfortable if I knew that the open source software that I downloaded for free gave me the source code and so I can fix or modify the application as per my needs. But I would be more comfortable thinking that the paid software I use from the vendor is going to be in the industry for long and many brains are working on the software to ensure that it runs smoothly and has improvised functionality with every new release.

All this does not mean that open source is a big NO-NO. Open source does provide the benefit of using someone’s work in your own and then redistributing your work in similar fashion so that someone else can get benefit from your work. This way the community can make progress using a collaborative manner of production. Most of the time this needs to be shared and released under GPL (General Public License) which means that you cannot charge others for a license and you must make your own source code available to the public. This makes it IP (Intellectual Property) unfriendly. Many a times this code will not even make its way to the main stream and will simply end up as an individual company’s product. This leaves a question in mind – will everything in the long run be compatible or provide interoperability? Will my investment in one segment ensure returns in the other or vice-versa? Who’s going to provide updates and security patches to fix vulnerabilities? Who’s going to invest billions to innovate? From my perspective, it’s not a fight between open source and commercial software. It’s all about open standards versus proprietary standards.

I feel we should have more specific and beneficial standards that are not vendor specific or not vendor dictated because ultimately it is the interoperability that counts at the end of the day. If open source software fits your environment and gets the work done in terms of costs, features, support or maintenance – all’s well. But if you are putting security, compliance, performance, upgrades and scalability before everything else then proprietary software designed with open standards in mind is your choice. We can even extend this further and run a combination of both – it’s our choice.

What do you think? Let me know.

Until the knock of environment friendly concepts and public concerns, traditional datacenters had fixed their eye on maximum uptime. Times have changed and datacenters are concentrating on conserving energy. Here are some tips that can contribute towards building eco-friendly datacenters.

Recently, I was involved in redesigning our company’s datacenter to consolidate hardware and systems running at two different physical locations. This was planned not as a “marketing” or “great idea!” move, but in fact it was a pragmatic business decision that helped the company to save money each year. I would like to share here my key learning’s that helped us make decisions to work towards a Greener Datacenter.

Evaluate Current Infrastructure: This is the first step where analysis needs to be done on the energy bills, current consumptions, and how much is being utilized by the servers, air conditioning systems, cooling fans and other devices like routers and switches in the datacenter. This helps in creating a baseline for calculating the Return-On-Investment after the changes are actually put in to implementation. Here is a link to couple of Power Calculators:

India – http://www.indiaenergyportal.org/
UK – http://www.ukpower.co.uk/running-costs-elec.asp
USA – http://www.42u.com/efficiency/energy-efficiency-calculator.htm

Rack Placements: This is an important aspect while designing the datacenter floor plan. Plan the layout of the equipments and structural components in a hot aisle and cold aisle rack layout method as recommended by Green Grid. The basic principle is to maximize the separation between the exhaust air flow and intake air flow. The figure below would give a better idea of the basic design principle.

The benefit of this arrangement is that it affects the required air delivery temperature that must be used to equalize the temperature throughout the room.

Need Analysis: Evaluate what is required out of the datacenter, not just what is needed now but maybe two or three years down the line. This helps in estimating the room for growth and helps avoid costly mistakes which fail to allow enough headroom for IT growth.

Eliminate: Look at every single area where you can eliminate, reduce or consolidate current hardware and equipments. For example, if you have two 24 port switches using only 18 connections consolidate them and shut down one of them. This does not have much impact in small organizations, but when we talk about companies which have lot of equipments in their datacenter, I’m sure figures would be impressive. Moreover, while calculating the ROI also consider the capital equipment and ongoing operational cost savings.

Consolidation of Servers: Identify areas where servers can be consolidated on to one single piece of hardware rather than running multiple physical boxes. By running multiple platforms and applications on a single server with many virtual servers can help reduce the amount of physical boxes. This in turn would help save the power consumption of not only the hardware box but also affect the cooling systems energy consumption.

Usage Patterns: There is a misconception that servers in a datacenter should be available and performing 24×7 regardless of the usage levels. In fact, identify and list out how all different parts of the network are being utilized and what are the peak hours for bandwidth, processing power and storage to handle peak network traffic. Set the power savings features on the hardware or OS level such that it helps in saving power. Allow the servers to sleep when they are not being utilized and explore the Wake-On-LAN technology.

Rated Appliances: Use Energy Star rated appliances, specially the air-conditioning or cooling systems. Devices carrying the Energy Star logo tend to save about 20% to 30% of energy on an average and the specifications differ with each item. Keep an eye on this point as it is estimated that EPA (Environmental Protection Agency) will announce and enforce the ratings for datacenter appliances and servers by 2010. At times, devices with Energy Star rating come costly but it is worth investing that marginal amount during the initial purchase. The overall running costs of power consumption would be much higher if the devices are not power efficient and you will run spending more on the energy costs in the long run.

Monitoring: It’s important to keep a track on the uptime and usage patterns of the datacenter. It basically allows the IT managers to accurately measure the effectiveness of the changes. Moreover, it also assists the top level management to decide on green energy solutions to reduce or eliminate their carbon footprints.

Remember, Green is just not a color – it’s more than that.

Do feel free to add in your comments, ideas and experiences. They would be much appreciated by all those who are planning to make computing energy efficient.

One of the features in ISA Server 2006 is the ability to block traffic based on URL or Domain name. This means that traffic can be blocked for a particular website from ISA Server without disrupting the general Internet access rule.

I’ve compiled some Domain Name Sets and URL Sets from the Internet and zipped them for easy availability for ISA administrators. Download the ZIP file and extract it. Under Network Objects in the Toolbox tab, right click URL Sets and click Import. Choose a single XML file from the unzipped folder of URLs. Once you have imported all XMLs, follow the same procedure for Domain Name Sets.

The next step is to create a rule which denies traffic to the websites which are listed in the XML files that we imported. Start by creating a new rule. I’ve named my rule as “Block Custom Sites”.

In the Access Rule, choose “Deny”.

Under protocols, choose HTTP and HTTPS.

Under Sources, choose Internal and VPN Clients.

Under Destinations, choose the XML lists that we imported. You can add multiple XML files.

Remember to shift the rule that we created to the top of all rules and we are done.

All those System Administrators running their Exchange Server 2007 on a Windows 2003 (64-bit) server have the previlage to get access to the traditional NTBackup utility to backup their exchange stores. But those who are running their Exchange Server 2007 on Windows 2008 (64-bit) must have faced the dilema of which tool to use to backup their mail stores. This is because Windows Server Backup in Windows Server 2008 no longer supports Exchange-aware backups or restores. In order to back up and restore Exchange Server 2007 Service Pack 1 (SP1) on Windows Server 2008, you must use an Exchange-aware application that supports the Volume Shadow Copy Service (VSS) writer for Exchange 2007, such as Microsoft System Center Data Protection Manager, a third-party Exchange-aware VSS-based application, or a third-party Exchange-aware application that uses the streaming backup APIs locally on the Exchange server to make a backup locally on the Exchange server.

What does all this mean? Nothing but shelling out extra $$$ to invest in a backup solution unless you are interested in a very simple backup solution to assure that the logs get truncated and you have backup copies which you can restore. Here is a basic hack which can help one save a couple of $$$.

Step #1

Get access to a Windows 2003 R2 Server which is running a 64-bit OS. Assuming that you are not running your production Exchange Server 2007 in 32-bit environment, we require the 64-bit version library files. If you do not have a one handy, you can get the source files and then expand the DLL’s or install one in a virtual environment.

Step #2

Copy the ntbackup.exe, ntmsapi.dll, and vssapi.dll files from C:\windows\system32 into a new folder on your Exchange Server 2007 running on Windows 2008. I created a folder called NTBackup under Program Files and placed them there.

NT Backup Exchange 2007

Step # 3

Right click Ntbackup.exe and choose “Run as administrator”.

Your should see the Exchange Server Information Store as part of your backup wizard.

Select the Information Stores that you want to backup. Choose the Backup media or filename and the path where you want to store your backups. Once you have finished the configuration, click on Start Backup and the exchange store backups should start.

Once the backup is complete, you should see all the transaction logs getting truncated (hope circular logging is not enabled) and also the mail store would be stamped with the Last Full Backup date and time stamp.

One thing to note though: you cannot back up a storage group in a Standby Continuous Replication (SCR) environment. Backups of storage group copies are available for Local Continuous Replication (LCR) or Cluster Continuous Replication only.