Daily Archives: April 17, 2004

Gmail and the Internet OS

Posted on by

Tim O’Reilly puts Google’s Gmail in a wider context:

Gmail is fascinating to me as a watershed event in the evolution of the internet. In a brilliant Copernican stroke, gmail turns everything on its head, rejecting the personal computer as the center of the computing universe, instead recognizing that applications revolve around the network as the planets revolve around the Sun. But Google and gmail go even further, making the network itself disappear into the universal virtual computer, the internet as operating system.

I’ve been dreaming this dream for years. At my conference on peer-to-peer networking, web services, and distributed computation back in 2001, Clay Shirky, reflecting on “Lessons from Napster”, retold the old story about Thomas J. Watson, founder of the modern IBM. “I see no reason for more than five of these machines in the world,” Watson is reputed to have said. “We now know that he was wrong,” Clay went on. The audience laughed knowingly, thinking of the hundreds of millions, if not billions, of computers deployed worldwide. But then Clay delivered his punch line: “We now know that he overstated the number by four.”

Pioneers like Google are remaking the computing industry before our eyes. Google of course isn’t one computer — it’s a hundred thousand computers, by report — but to the user, it appears as one. Our personal computers, our phones, and even our cars, increasingly need to be thought of as access and local storage devices. The services that matter are all going to run on the global virtual computer that the internet is becoming.

Until I heard about gmail, I was convinced that the future “internet operating system” would have the same characteristics as Linux and the Internet. That is, it would be a network-oriented operating system, consisting of what David Weinberger calls “small pieces loosely joined” (or more recently and more cogently, a “world of ends”). I saw this as an alternative to operating systems that work on the “one ring to rule them all principle” — a monolithic architecture where the application space is inextricably linked with the operating system control layers. But gmail, in some sense, shows us that once storage and bandwidth become cheap enough, a more tightly coupled, centralized architecture is a real alternative, even on the internet. (I have to confess that was one of the wake up calls to me in Rich Skrenta’s piece, linked to above.)

But in the end, I believe that the world we’re building is too complex for tight coupling to be the dominant paradigm. It will be a long time, if ever, before any one company is in control of enough programs and enough devices and enough data to start dictating to consumers and competitors what innovations will be allowed. We’re entering a period of renewed competition and innovation in the computer industy, a period that will utterly transform the technology world we know today.

I love Dave Stutz’s phrase, “software above the level of a single device.” We’re used to thinking of software as something that runs on the machine in front of us, its complex dance hidden by the blank metal and plastic of the hardware that houses it. But now, computers are everywhere, and each dance has many partners, a whirling exchange of data that will be made visible when and where we want it. It’s not the machine or even the software that matters, it’s the information and services that travel over the hardware and software “wires.” Gmail’s introduction of large amounts of free online storage for application data is an important next step in freeing us from the shackles of the desktop.

This isn’t to say that there aren’t important issues raised by the internet paradigm shift. The big question to me isn’t privacy, or control over software APIs, it’s who will own the data. What’s critical is that gmail makes a commitment to data migration capabilities, so the service isn’t a one way door to the future. I want to be able to switch to alternate providers if the competition makes a better offer. The critical enabler is going to be the ability to extract my data and connections so that I can work with them on multiple devices, for example, syncing my laptop or phone with my gmail account rather than having to work only in a tethered fashion. I understand why gmail doesn’t offer this feature now, but it’s going to be essential in the long term.

More on A9 – A History Server

Posted on by

John Battelle provides more insights into Amazon’s A9 search engine:

To me, the core feature that makes A9 interesting is what Udi Manber calls a “history server” – the technologies behind A9’s search history and personalization features. Having your entire search and click history, and if you use the Toolbar, your entire browsing history as well, available on a server side application opens up all sorts of new approaches to solving search, research, and recall problems. Combining that history with what Amazon already knows about you (no, A9 does not do that…yet) creates even more powerful possibilities.

What gets me thinking is that for those who commit to A9 as a search solution, new and continuous improvements in search are likely to be hacked up, based on the fact that the personalized history can be analyzed and leveraged. For example, Gary and others have noted that the service does not allow you to keyword search within your searches, and display, for example, just those pages you’ve browsed in the past. I’d wager Giants tickets that will be in the feature set by the end of the year.

A minor example of the power of the history server: when you repeat a search, A9 will show you what links have changed and what links you’ve clicked on before. This might seem like a minor deal, but it’s a pretty effortless feature for A9 to serve up. Imagine what else might be done with the history server. If you can imagine it, you can probably do it – again, I’ll wager that Amazon will figure out a way to make the A9 interface API friendly, so that its platform developers can cook up even greater feats.

Other useful links:
– Business 2.0: Joh Battelle’s interview with Amazon’s Udi Manber
– Rex: “To me, A9 is not designed as an Internet search engine, but as a knowledge-searching tool to end all knowledge searching tools…..As you look for information, Amazon will provide you the results that ‘people like you’ have found most helpful when searching for the same information, product, place, answer, etc..I don’t think Amazon wants to compete with Google. Google admitted recently that it was a content business. Amazon has no such designs. Amazon, rather, wants to connect you with something you can purchase.”
– Christian Lindholm says that Search has turned to Find with A9

We are at an exciting time in the evolution of the Internet – innovation is happening at many companies and we bloggers can discuss these events in real-time, matching our wits against each other and the developers at the companies.

MailServ: FCB-Ulka Case Study

Posted on by

CXOtoday.com has a case study about FCB Ulka’s deployment of our Linux-based messaging solution (Emergic MailServ):

The FCB-Ulka group comprising of FCB-Ulka Advertising and Interface Communications (offering marketing and communication consultancy services), has recently switched its messaging solution from Sendmail to Emergic Mailserv – a Linux-based messaging and security solution.

Developed by Netcore Solutions Pvt Ltd., the highly customized distributed solution has been deployed at the groups seven locations across six cities in India.

It all started in 1999, when FCB-Ulka, part of the FCB Group and Interpublic Group of Companies, decided to switch from MS Exchange to the Linux-based Sendmail. Deployed by Ashtech Infotech Ltd., the solution worked well over a period, but eventually with a rise in scope and usage, the drawbacks far outweighed the pros.

Ritu Madbhavi, systems director, FCB Ulka Advertising, outlined some of the issues the company encountered, “Administration difficulties such as tedious process of adding and deleting new users, lack of simple interface for basic functions like auto responder and disclaimer arose with Sendmail.”

To add to their woes, despite developing new scripts with the latest versions of Perl, they were incompatible with the older version of Linux.

Obsolesce was another issue the media house had to contend with, as the system was built on Red Hat Linux 5.0, which used fetchmail. Remote access was provided via dialup, but it was limited to only one person being able to access mail at any given time due to a single modem with a telephone line on the server.

Moreover, the mailing architecture was such that one of the locations in Mumbai acted as a hub. So a breakdown of the server in that location ended up affecting the entire companys email structure.

Security too reared its head, as the system was a basic messaging with no firewalls, and the sole security provided for by the routers. Even monitoring of internet access was not possible as the server was equipped with only the basic proxy services.

Explained Madbhavi, FCB-Ulka was now looking for something beyond messaging and we evaluated various solutions. Having run a Linux based mailing solution, the team at FCB-Ulka knew exactly what they wanted in the new mailing system. The requirements were crystallized and the only company that came close to satisfy our needs was Netcore, which in turn implemented its flagship Emergic Mailserv.

The basic functionalities implemented in the messaging area was a standard compliant mail server providing IMAP, POP, SMTP, LDAP services, availability of webmail allowing users to access when they are away from the desktop or when they are traveling.

In the security space, the mail is now handled from a web-based server managed 24×7 by Netcore, an anti-spam that detects junk based on the content, tags them and then blocks it at the ISP, server and user level. Unlike the previous system, which had a limitation of one user only, the new one can handle any number of users. The advance proxy enables the system administrator to block sites based on various criteria such as domain name, content, etc. Also, the authentication based proxy controls the browsing schedule deciding which user is allowed to browse when.

According to Kalpit Jain, CTO of Netcore, The solution that runs only on Linux, has an anti-virus application, armed with a firewall, enhanced with load balancing features apart from bandwith management and monitoring. The mail server has been placed in the DMZ (DeMilitarized Zone)- a server zone, which is accessible to the public as well as to the private network. In other words, it separates the internal network from the outside world.

Commenting about the inherent benefits, Madbhavi cited reliability, flexibility coupled with ease of administration and simplicity. However, acknowledged Madbhavi, Netcore was able to execute a high degree of customization primarily because the solution was built on open source.

Emergic Mailserv has been built with a judicious mix of certain open source components alongwith internally developed software, affirmed Jain.

Spelling the minimum technical requirements, Jain stated that for 25 users a P III with a 64 MB RAM with 20 GB hard disk is needed.

Founded in 1998, Netcore Solutions Pvt Ltd. is an enterprise solutions company, focused on messaging, collaboration, and security software offering a range of enterprise products and hosted services. Among its several corporate customers using this solution are IDBI Bank and Raymonds.