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.
On Fri, Sep 12, 2008 at 5:53 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 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.
> From: Khazret Sapenov <sape...@gmail.com> > Subject: e-government in cloud > To: cloud-computing@googlegroups.com > Date: Friday, September 12, 2008, 5:53 PM > 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.
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.
On Fri, Sep 12, 2008 at 6:18 PM, Khazret Sapenov <sape...@gmail.com> wrote:
> 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.
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 ?
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 :)
On Fri, Sep 12, 2008 at 6:34 PM, Khazret Sapenov <sape...@gmail.com> wrote:
> 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 ?
Khazret Sapenov 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 > 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.
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
On Fri, Sep 12, 2008 at 6:43 PM, Jim Starkey <jstar...@nimbusdb.com> wrote:
> Khazret Sapenov 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 > > 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. > 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.
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.
"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.
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.
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.
> 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.
> > 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.
> 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.
> > > 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.
> > 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.
> > 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.
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.
________________________________
From: cloud-computing@googlegroups.com
[mailto:cloud-computing@googlegroups.com] On Behalf Of Khazret Sapenov
Sent: Friday, September 12, 2008 5:53 PM
To: cloud-computing@googlegroups.com
Subject: e-government in cloud
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
<html>
<body>
<p>***********************************************************************< /p>
<p>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.</p>
<p>***********************************************************************< /p>
</body>
</html>
On Fri, Sep 12, 2008 at 3:52 PM, Tim Freeman <tfree...@mcs.anl.gov> wrote: > On Fri, 12 Sep 2008 17:53:27 -0400 > "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. > 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.
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
> 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.
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).
On Tue, Sep 16, 2008 at 3:50 AM, Daniele <daniele.ta...@gmail.com> wrote:
> 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 > > 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.
|
