e-government in cloud

35 views
Skip to first unread message

Khazret Sapenov

unread,
Sep 12, 2008, 5:53:27 PM9/12/08
to cloud-c...@googlegroups.com
Not an actual use case yet, but I see that electronic voting (e.g. US presidential elections) would be a good use case for cloud computing, given technology allows eliminate electoral fraud and provides reliable audit trail.
Once in four years, it would take to process about 305 millions of voting ballots (as of today according to US Census http://www.census.gov) within short period of time. 

It might be supplementary to existing federal election system and later they may decide to transition to this model completely. Introduction of such technology should reduce processing time and costs (does $10,000 budget sound good enough?:) ). Shaving off TCO of electoral equipment would always please ministry of finance. 

Another use case in my mind would be optimization of USPS, that might be beneficial for both government, businesses and cloud providers, but I'll leave it here, until I'm ready with specifics.

cheers,
Khazret Sapenov 

Andrew Badera

unread,
Sep 12, 2008, 5:55:11 PM9/12/08
to cloud-c...@googlegroups.com
Why does the cloud offer anything that any third party datacenter cannot?

--Andy Badera

Khazret Sapenov

unread,
Sep 12, 2008, 6:00:53 PM9/12/08
to cloud-c...@googlegroups.com


On Fri, Sep 12, 2008 at 5:55 PM, Andrew Badera <and...@badera.us> wrote:
Why does the cloud offer anything that any third party datacenter cannot?

--Andy Badera


Scalability, lower TCO, more available capacity, than conventional datacenters can't guarantee, you name it...

ZUL KAGALWALLA

unread,
Sep 12, 2008, 6:00:04 PM9/12/08
to cloud-c...@googlegroups.com

Another one would be turbotax etc.

--- On Fri, 9/12/08, Khazret Sapenov <sap...@gmail.com> wrote:

Andrew Badera

unread,
Sep 12, 2008, 6:04:16 PM9/12/08
to cloud-c...@googlegroups.com
I think you're misunderstanding the problem of elections.

You have a pretty known quantity of demand over a very quantifiable, often short, period of time.

You could lease a datacenter for a day. It's not a matter of a year-long or years-long engagement.

You really think TCO on anything online is going to be lower than anything offline? Especially when you consider admin and auditing?

Khazret Sapenov

unread,
Sep 12, 2008, 6:18:49 PM9/12/08
to cloud-c...@googlegroups.com


On Fri, Sep 12, 2008 at 6:04 PM, Andrew Badera <and...@badera.us> wrote:
I think you're misunderstanding the problem of elections.

You have a pretty known quantity of demand over a very quantifiable, often short, period of time.

You could lease a datacenter for a day. It's not a matter of a year-long or years-long engagement.

You really think TCO on anything online is going to be lower than anything offline? Especially when you consider admin and auditing?


Well, I think you need more than a day to deploy and test solution, perhaps a month or two would be safe enough. Also you'd need several locations for higher availability.

TCO would always be lower with cloud, since you scale up/down dynamically versus fixed size capacity and charged accordingly.

Andrew Badera

unread,
Sep 12, 2008, 6:21:31 PM9/12/08
to cloud-c...@googlegroups.com
But election needs aren't that dynamic ... they're pretty easily defined as static :)

Khazret Sapenov

unread,
Sep 12, 2008, 6:34:55 PM9/12/08
to cloud-c...@googlegroups.com


On Fri, Sep 12, 2008 at 6:21 PM, Andrew Badera <and...@badera.us> wrote:
But election needs aren't that dynamic ... they're pretty easily defined as static :)


What I want to say is, that scenario with unnecessary over-provisioning won't please any CFO :)
I assume votings are spread across time period, so you dynamically provision/deprovision capacity according to current load (or maybe I'm mistaken and all 300 mln make their choice simultaneously within 1 minute?). Example from real life would be using 5 buses (40 seats each) to serve party with known number of guests. If you can do it for less money, why overpay?
Does it make sense ? 

Andrew Badera

unread,
Sep 12, 2008, 6:41:57 PM9/12/08
to cloud-c...@googlegroups.com
Again, you're misunderstanding the election problem. You're creating a situation that does not currently exist. 300M people can't vote at the same time, and we don't currently have any means, resource provision/deprovision aside, from allowing all 300M to vote at their whim :)

Jim Starkey

unread,
Sep 12, 2008, 6:43:31 PM9/12/08
to cloud-c...@googlegroups.com
Ignoring that a) voting policy is set by the states, not the Feds, b)
optical mark balloting is rapidly replacing electronic voting for all of
the right reasons, and c) the electoral college makes a central audit
meaningless, who in their right mind would trust an audit performed by a
private company with the government unable to even verify the physical
integrity of the data, let alone monitor the tons and tons of the
proprietary software used to manage the cloud? I trust Amazon to sell
me books and not give away my credit card numbers, but select a
president? You've got to be kidding!

Also, if there are 305 million ballets, we're already in deep trouble
since only about half the population is of voting age.

If you didn't like the NPR report on disappeared data, you're really
going to dislike what they have to say about this.

--
Jim Starkey
President, NimbusDB, Inc.
978 526-1376

Khazret Sapenov

unread,
Sep 12, 2008, 6:56:19 PM9/12/08
to cloud-c...@googlegroups.com
Ah, my bad, I was looking for voting quantity, but was distracted somehow and taken entire population, still 150 mln looks like a lot. :)

I'm not an election expert (not even an American), just suggested potential use of cloud to shave off a couple of bucks. Looks like it doesn't fit.

  

Tim Freeman

unread,
Sep 12, 2008, 6:52:16 PM9/12/08
to cloud-c...@googlegroups.com
On Fri, 12 Sep 2008 17:53:27 -0400
"Khazret Sapenov" <sap...@gmail.com> wrote:

> Not an actual use case yet, but I see that electronic voting (e.g. US
> presidential elections) would be a good use case for cloud computing, given
> technology allows eliminate electoral fraud and provides reliable audit
> trail.

I hope it's never a use case. How could it ever guarantee no fraud or a
reliable audit trail?

An honest vote is one of THE most important things we need to be paranoid
about. The payoff for cracking is very high and the results possibly
devastating for the entire world.

It's Friday, so I'm feeling free to troll a little bit. :-) This rant applies
to e-voting too.

=====

We are told that "election fraud" is a much bigger problem than "voter
fraud" (i.e., someone wanting to infiltrate voting systems/procedures and do
something nasty to the count is a much bigger issue than an army of individuals
pretending to be someone else etc.) which intuitively makes sense.

Having a few key (and potentially interconnected) targets to inflict damage
upon makes this even easier, no matter how secure of a system you think you
have.

Cloud offload or online voting would require the internet to be involved: ugh.
Even the most secure endpoints in the world can have trouble on the net, it's a
disaster waiting to happen. I'd be fine with this aspect of it if a diety
designed and wrote the software and ran the network and datacenters, but I
don't think one is available. We all know there are plenty of people who are
"smarter" than we are, and I don't think it's a good idea to assume they are
all not criminals.

And it would mean that the data is potentially processed and stored across
several 3rd party sites: ugh. The prize is so great that I can imagine
"sleeper" agents being deployed, employees that are sent years in advance to
work for cloud providers :-)

And paper ballot records are essential, records that are geographically
distributed across the country to make it virtually impossible for someone to
alter systematically. Computers make things relatively easy to alter
systematically.

Bruce Schneier on the subject:

http://www.schneier.com/essay-039.html

http://www.schneier.com/blog/archives/2006/11/more_on_electro.html

Risk vs. reward is a good predictor of human behavior, one must always allow
for human error (in code or operations), and one must always assume there are
far smarter people out there up to no good.

So when Bruce says "simplicity is the key" it makes a lot of sense to me. The
security risks of a simple system are far easier to suss out and address.

Tim

Khazret Sapenov

unread,
Sep 12, 2008, 7:11:37 PM9/12/08
to cloud-c...@googlegroups.com
On Fri, Sep 12, 2008 at 6:52 PM, Tim Freeman <tfre...@mcs.anl.gov> wrote:

On Fri, 12 Sep 2008 17:53:27 -0400
"Khazret Sapenov" <sap...@gmail.com> wrote:

> Not an actual use case yet, but I see that electronic voting (e.g. US
> presidential elections) would be a good use case for cloud computing, given
> technology allows eliminate electoral fraud and provides reliable audit
> trail.

I hope it's never a use case.  How could it ever guarantee no fraud or a
reliable audit trail?

An honest vote is one of THE most important things we need to be paranoid
about.  The payoff for cracking is very high and the results possibly
devastating for the entire world.

Well, I assume there's only one outcome out of possible 2-3 and probability of second Hitler to become a president is pretty low (both candidates should have satisfy some minimal level). So it'll devastate only a half of voters in one country. :)   
 

And it would mean that the data is potentially processed and stored across
several 3rd party sites: ugh.  The prize is so great that I can imagine
"sleeper" agents being deployed, employees that are sent years in advance to
work for cloud providers :-)

No too convincing :)
Nuclear reactors are really more dangerous and managed by third party and audited by government.
 

And paper ballot records are essential, records that are geographically
distributed across the country to make it virtually impossible for someone to
alter systematically.  Computers make things relatively easy to alter
systematically.

Bruce Schneier on the subject:

   http://www.schneier.com/essay-039.html

   http://www.schneier.com/blog/archives/2006/11/more_on_electro.html

Risk vs. reward is a good predictor of human behavior, one must always allow
for human error (in code or operations), and one must always assume there are
far smarter people out there up to no good.

So when Bruce says "simplicity is the key" it makes a lot of sense to me.  The
security risks of a simple system are far easier to suss out and address.


Agreed 101% 


Jim Starkey

unread,
Sep 12, 2008, 8:18:00 PM9/12/08
to cloud-c...@googlegroups.com
Khazret Sapenov wrote:
>
> Ah, my bad, I was looking for voting quantity, but was distracted
> somehow and taken entire population, still 150 mln looks like a lot. :)
>

Hey, I'm from Chicago. A voter turnout that exceeds the population
isn't all that unusual.

Tim Freeman

unread,
Sep 12, 2008, 8:27:23 PM9/12/08
to cloud-c...@googlegroups.com, sap...@gmail.com
On Fri, 12 Sep 2008 19:11:37 -0400
"Khazret Sapenov" <sap...@gmail.com> wrote:

> On Fri, Sep 12, 2008 at 6:52 PM, Tim Freeman <tfre...@mcs.anl.gov> wrote:
>
> >
> > On Fri, 12 Sep 2008 17:53:27 -0400
> > "Khazret Sapenov" <sap...@gmail.com> wrote:
> >
> > > Not an actual use case yet, but I see that electronic voting (e.g. US
> > > presidential elections) would be a good use case for cloud computing,
> > given
> > > technology allows eliminate electoral fraud and provides reliable audit
> > > trail.
> >
> > I hope it's never a use case. How could it ever guarantee no fraud or a
> > reliable audit trail?
> >
> > An honest vote is one of THE most important things we need to be paranoid
> > about. The payoff for cracking is very high and the results possibly
> > devastating for the entire world.
> >
>
> Well, I assume there's only one outcome out of possible 2-3 and probability
> of second Hitler to become a president is pretty low (both candidates should
> have satisfy some minimal level).
> So it'll devastate only a half of voters in one country. :)

Heh, I'm deleting comment left and right here, trying to stay on track and
polite: let me at least say that there are some perfect examples of bad
judgement yielding disastrous consequences in recent years -- and definitely
not just limited to the country of the vote.

> >
> > And it would mean that the data is potentially processed and stored across
> > several 3rd party sites: ugh. The prize is so great that I can imagine
> > "sleeper" agents being deployed, employees that are sent years in advance
> > to
> > work for cloud providers :-)
> >
>
> No too convincing :)
> Nuclear reactors are really more dangerous and managed by third party and
> audited by government.

I'm assuming you're saying that since nothing purposefully bad has happened with
nuclear reactors, they are secure?

I would argue that the reason we have not seen problems there is a (relative)
lack of motivation to do such horrific things.

The rewards for throwing an election are far higher and more commonly desired
than that. The thirst for power, money, and fame can be insatiable and is very
common -- but for senseless death and destruction? Not common at all
(relatively speaking).

And the more people that would be motivated, the more there are that would
actually go for it. And so there's a higher likelihood that someone would
succeed in my opinion. In the case of elections and political power struggles
in general, deception and crime has happened a lot (provably).

The candidates themselves do not even need to be involved in such a conspiracy
for power and money to come to those who would throw an election.

Tim

Forman, Mark A

unread,
Sep 12, 2008, 8:42:04 PM9/12/08
to cloud-c...@googlegroups.com, Smith, Brian A

As the first U.S. E-Govt Administrator, I am very intrigued by applications of Cloud Computing to government.  The best use case for this would be electronic filings, especially quarterly Tax payments and annual Tax returns.  However, business gateways and citizen service portals that cut across levels of government would also be good.  Elections would not, for a number of reasons.

 


***********************************************************************

The information in this email is confidential and may be legally privileged. It is intended solely for the addressee. Access to this email by anyone else is unauthorized. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited and may be unlawful. When addressed to our clients any opinions or advice contained in this email are subject to the terms and conditions expressed in the governing KPMG client engagement letter.

***********************************************************************

Dan Kearns

unread,
Sep 13, 2008, 12:38:02 AM9/13/08
to cloud-c...@googlegroups.com
On Fri, Sep 12, 2008 at 3:52 PM, Tim Freeman <tfre...@mcs.anl.gov> wrote:
On Fri, 12 Sep 2008 17:53:27 -0400
"Khazret Sapenov" <sap...@gmail.com> wrote:
> Not an actual use case yet, but I see that electronic voting (e.g. US
> presidential elections) would be a good use case for cloud computing, given
> technology allows eliminate electoral fraud and provides reliable audit
> trail.
I hope it's never a use case.  How could it ever guarantee no fraud or a
reliable audit trail?

I think It'd be pretty cool to have the raw ballot info available and turn all the map-reduce fanatics loose on it. I suspect there would be enough interesting ways to process the data that a) it could suck up an impressive number of cpu hours and b) it could suggest a bunch of interesting statistical properties, including fraud detection.

-d

Daniele

unread,
Sep 16, 2008, 3:50:08 AM9/16/08
to Cloud Computing

At CNIPA (the Italian eGovernment-supporting technical body) we are
just organizing a series of seminars on cloud computing with industry
and consultancies. My understanding is that electronic voting will not
be a viable use case for a number of reasons which have little to do
with the infrastructure or the paradigm that might be used.

Much better eGovernment use case could be a "my government" portal,
where people could exchange official communications with the
administration, update their filings and perform payments after being
digitally identified. A cloud infrastructure could be the best choice
to experiment with a "zero-latency government", where back-office
interactions are not visible to citizens.

There are a lot of things that should be analysed and piloted before,
including legislative and privacy issues. The effort could be
extremely rewarding, though.



On Sep 12, 11:53 pm, "Khazret Sapenov" <sape...@gmail.com> wrote:
> Not an actual use case yet, but I see that electronic voting (e.g. US
> presidential elections) would be a good use case for cloud computing, given
> technology allows eliminate electoral fraud and provides reliable audit
> trail.
> Once in four years, it would take to process about 305 millions of voting
> ballots (as of today according to US Censushttp://www.census.gov) within

Khazret Sapenov

unread,
Sep 16, 2008, 10:09:37 AM9/16/08
to cloud-c...@googlegroups.com
Perhaps initially governments are better off with 'government cloud', where different agencies might run pilot programs to test feasibility, so it will be easier to comply to certain criterium for privacy and security. If results are negative, it might be decommissioned and sold to third party.
I would call this approach 'trans-cloud'.

As for potential use cases, off top my head, I'd look into e-learning as well. Generally we are looking for pattern, where processes have periodic spikes (say, weekly draws lottery draws).

cheers,
Khazret Sapenov
Reply all
Reply to author
Forward
0 new messages