“So let me take a bit of time to write a short narrative
of the story of Sahana .. told from my memory and perspective. My
apologies if I missed any key people (I probably did :(..) and for any
other mistakes in my narrative. I’m on a long flight so I don’t have net
access to check my old blogs to validate the dates. If anyone is
interested you can read the blogs I wrote during the tsunami to see the
gory details of the really early days; Google will help you find those.
Sunday, December 26th, 2004. Tsunami hits Indonesia, Sri Lanka and many
other Asian countries. In the first week of the tunami, 1m people (or 5%
of our population) was homeless. 2/3rds of Sri Lanka’s coast was affected
in some way. Later on we find that nearly 40,000 of our people have died.
Tuesday, December 28th, 2004. Many different organizations in Sri Lanka
start efforts to write various bits of software to help manage the
disaster. (This bit of the story was repeated in other countries- India,
Indonesia, Thailand etc..)
Wednesday, December 29th, 2004. Many of these folks get together at the
ICT Agency in Narahenpita, Sri Lanka to discuss ways of putting the
software all together to make it easier to manage the situation. That nite
I called the US Federal Emergency Management Agency (FEMA)’s CIOs office
(after finding the phone number in a powerpoint presentation he had done
proposing developing a disaster management software system) and asked for
whatever software they had. I was told that FEMA had no software that
could help .. they only had software that was used to cut checks to people
after hurricanes.
In the 3-4 weeks that followed, many many individuals, universities and
software companies and Sri Lanka Telecom contributed to what became known
as Sahana. Amongst the IT companies, Virtusa was the leading contributor
with more than 75 of their engineering staff helping at some time or the
other. While most contributors to the initial effort were from Sri Lanka,
I’d be remiss if I didn’t mention the urgent support we got from folks at
Tigris.org (which we didn’t end up using) and later SourceForge. We
desperately needed a code repository and other infra (like mailing lists)
and these folks willingly and urgently came out of their holiday slumber
and got everything that we needed. Special mention also needs to go IBM’s
Crisis Response Team lead by Brent Woodworth, who were then regular
visitors to Sri Lanka. From day 1 that entire team supported, encouraged
and cheered on the Sahana effort. In fact a good part of the initial
development was done on 15 notebooks that IBM donated within a week or so
of the tsunami.
This joint effort was organized and managed by the Lanka Software
Foundation. In the early days we had a 24×7 operation and the first bits
of software went into production use in about a week. Over time more and
more capabilities were added and used in various ways. After about 3
months this initial phase was completed and the software and its
deployment reached a certain level of equilibrium.
In the meantime, it became clear to us that there was a huge hole in the
world of disaster management software. The state of the art that the UN
team that came to Sri Lanka with was a system called SUMA- something
written on FoxPRO. (Anyone remember FoxPRO? Yes, that was the
pre-relational desktop database system from Microsoft!) IBM had some stuff
based on Lotus Notes but it wasn’t easily deployable, scalable and, most
importantly, didn’t embrace the Web. The tsunami gave us a unique
opportunity to look at disaster management in the modern world: even
though there was sooo much death and damage, the communication network was
in tact. Cell phones worked. The IP networks worked. Land-lines worked. A
modern disaster management system must work in a connected environment ..
and if communication has indeed failed (as often happens in earthquake
type disasters) its now quite easy to airdrop a box that sets up a local
communication network with a satellite uplink. Clearly, there was a huge
need for modern software that could live in this world and help first
responders and follow-up recovery folks be more effective at responding
and managing a disaster.
We were not going to let Sahana die; we decided we are going to make it
into something the world can reuse readily. “We” at the time was primarily
Jivaka Weeratunge, co-founder of LSF and its then volunteer COO, and myself.
Chamindra de Silva, who had been one of the original people from Virtusa
who started the people registry which became a key component of Sahana,
agreed to leave his job at Virtusa and take a 1-year position in LSF to
take Sahana forward if we could get the funding for it. Chamindra became
part of the “let’s take sahana forward” team.
On February 11th 2005 I wrote the following in a cover letter on the
proposal we submitted to Ms. Asa Heijne, First Secretary of the Swedish
Embassy in Colombo along with a proposal seeking Rs. 8.548m (approximately
$85k) in funding from the Swedish International Development Agency (SIDA)
to re-do Sahana:
“Further to our discussions in late January, enclosed please find a
proposal to further develop the Sahana Relief Management system into a
fully reusable, globalized relief management system. We believe the
potential global impact of such software will be tremendous and view this
as an opportunity to help the world at a time when the world is helping
Sri Lanka so willingly and widely.”
SIDA approved our proposal and Sahana phase II started with that funding
on August 1st 2005. I also want to acknowledge the contribution of
Per-Einar Troften in getting this funding- Per-Einar is in SIDA in Sweden.
He and Asa have (with the grant of $100k to start the Apache Axis2 project
and the Sahana grant) singlehandedly (two-handedly?) changed the role of
Sri Lanka in the FOSS world. If not for their trust in what LSF was
proposing to do Sahana wouldn’t exist in its current shape today.
I must also acknowledge LSF’s co-founder and then COO Jivaka Weeratunge-
he’s the one who helped manage LSF and make sure that we ran a superbly
tight and clear ship which made it easier for a funding agency to trust
us. Oh yeah, Jivaka was a total volunteer doing all of that, as is the
entire LSF Board. Jivaka was a key part of the strategy behind LSF overall
and then both and Sahana as we took them forward.
I think the following paragraphs we put in the proposal about why open
source was a critical component of disaster management may be useful for
folks to read:
“Very few countries and organizations today can afford to invest a lot of
resources in disaster management when there is no disaster present. While
this is obviously true of poor, developing nations, it is also true of
richer, developed countries as well because there are always higher
priority items that need the funding. Worse yet, even if there are some
national scale systems that may get deployed, it is very unlikely that
regional and local level systems will ever get deployed if they cost any
significant amount of resources.
Because no one is willing to pay for the software, no one is willing to
build it either. This is what we see in the world today – while disaster
management software is critically needed, there is no complete commercial
or non-commercial software solution that is widely available.
Going the open source way can address both these concerns. Using the open
source development model, it is possible to develop this software at a
much reduced cost compared to pure commercial development models. This is
true because while commercial entities are not willing to invest into
these systems, there are hundreds and thousands of well-meaning IT
professionals who are very happy to donate a few hours of effort to
helping build such systems. We are already seeing this with the nascent
Sahana project. Thus if there was a small team which was driving such a
project, then it is possible to get a lot of assistance from the global IT
community to make those systems truly exceptional.
Going with open source approaches can also greatly reduce the deployment
cost of this software in peace (i.e., non-disaster) times. The Sahana
system, for example, can be deployed on any PC with just a Linux LiveCD
(that is, a CD from which the entire system can be booted up and brought
on-line). Thus, not only is it possible to run this on commodity,
inexpensive hardware, it is in fact possible to not even have dedicated
hardware around – just take any office PC and make that the “disaster
management center”! In fact, that is how Sahana was first deployed in Sri
Lanka – on a borrowed PC. (Later it switched to running on a borrowed
server as the capacity requirements increased.)
Thus, open source is the natural way to providing disaster management
solutions.”
So that’s how Sahana Phase II was born.
LSF has managed the Sahana project (and charges 20% overhead on the human
resources part of the budget to do it .. a grand total of Rs. 1.008m or
around $10k for phase II) with the LSF board being the final authority for
how the Sahana team was deployed.
For those of you who know nothing about LSF- the board of LSF consists of
local software company senior executives (usually CEOs), heads of CS
departments of the 4 main public universities in Sri Lanka) and a few
other distinguished individuals. The board is not compensated and everyone
participates to help improve Sri Lanka’s position in FOSS- not for direct
commercial or personal benefit. We’re of course a non-profit organization
legally registered in Sri Lanka. LSF’s finances are annually audited by
Ernst & Young in Sri Lanka.
What LSF does is find the funding for and run projects like Sahana.
After the funding for Sahana from SIDA finished at the end of July 2006,
we’ve received a few additional grants .. with special thanks again to IBM
for both cash and significant hardware donations. Google also donated some
funds for LSF/Sahana.
Sahana has of course been a TREMENDOUS success. Kudos go to the core
development team (Chamindra, Pradeeper, Ravindra, Mifan and the rest of
the gang) for producing superb software, the committed bunch of folks on
the Sahana mailing lists (with special mention to Paul, Louiqa, Don and
Gav) and to the numerous others who have helped with developing Sahana,
deploying it or just talking about it. Special mention must go here to the
efforts of the IBM Crisis Response Team in deploying Sahana in numerous
disaster and pre-disaster situations. On the recognition side, the recent
FSF Award is clearly the high point, being the second recipient of that
after Wikipedia. The list of deployments of Sahana is absolutely
incredible .. and now includes both poor and rich (richest?) countries.
In this context the LSF Board started thinking last year about how to best
take Sahana forward and about the role of the LSF Board. We concluded that
the best thing to do was to hand over “reigns” of the LSF part of Sahana
to a new team of people who would be focused purely on making Sahana climb
as high as it can. In doing that, we CLEARLY separated the successful FOSS
project that Sahana is from the LSF managed work in developing and
deploying Sahana. In order to further the FOSS project of Sahana, we
created the Sahana Project Management Committee, modeled closely on the
Apache Software Foundation’s model. The FOSS project and the PMC are
purely community efforts- while we created the PMC, the future membership
of the PMC will be determined by the current PMC members. We bootstrapped
it and now its off on its own. Good luck!
The board has been appointed by LSF and will take overall charge of all
LSF activities related to Sahana, including budgets. Sahana Board members
are all volunteers and we’re extremely grateful for their willingness to
help take Sahana forward. LSF is the underlying legal authority for the
activities that the Sahana Board governs.
The specific roles and responsibilities of the community, PMC and the
Board were documented in an email I posted to this list earlier. See:
http://groups.yahoo.com/group/humanitarian-ict/message/2674
I hope this helps people understand how Sahana got started and how it has
evolved. Most importantly, I hope it makes clear the governance structure
of Sahana and its intrinsically open nature.
I personally drove the creation of this model (in close consultation with
a bunch of folks, esp. Chamindra, Louiqa, Paul, Don, Gav, Pradeeper and
more) and I used my 10+ years of experience with Apache and other open
source efforts to help create what I think is an absolutely open model.
That said, there’s always room for improving the structure and activities-
make constructive suggestions and I’m sure the community, the PMC, the
Sahana Board and the LSF Board will be willing listeners!”
