Using Outlook From Orbit

Pigskin-Referee writes with this excerpt from Office Watch: "On the Space Shuttle and the International Space Station they use Microsoft Outlook 2003, but not quite in the same way that us earthbound Earthlings do. The space shuttle Atlantis is orbiting the earth right now and the crew exchange emails with the ground a few times each day. Bandwidth is a constraint and you don't want the busy crewmembers bothered with spam or unnecessary messages so NASA has a special system in place. The crew use fairly standard laptops running Microsoft Outlook (currently Outlook 2003) with Exchange Server as the email host, but they don't link to the server using any of the standard methods."

Comments

Not simply webmail?

[Drethon]

Is it actually cheaper to upload all the e-mails in a burst instead of using a webmail system where only the mail the receiver wants to receive would be opened? Wouldn't work if they want to read offline I guess but the concern mentioned is bandwidth not connectivity.

Any mail experts comment?

Architecture?

[smitty777]

It's too bad the article didn't address the architecture behind all this. I would be curious to hear what kind of network they use, and what sort of relays (satellite?). If it is satellites, why is the bandwidth so low? (Hmmm... maybe they really should have made that ethernet cable just a little longer after all...)

Re:Architecture?

[Anonymous Coward]

I worked for the FAO, Flight Activity Office, developing applications about 15 yrs ago. These are the people who historically have provided email, fax and other document updates during missions to the astronauts for both shuttle and station missions.

At the time I was there, the data connection was 115 cps (no KB) connection. The IBM laptops were running WFW 3.11 with battery packs duct taped to the bottom. I think they took 4 or 5 up with each mission. The modems were highly specialized. Voice communications weren't encrypted, so email or fax was the most secure method of communication available. Also, the flight network in the FCR building was completely separate from the rest of the network at JSC for security concerns. That really isn't true, there was an outbound TTL limited connection for real-time data, but that didn't leave the campus. No inbound connections were allowed except through an "introduction" workstation or 1 computer in the FAQ back room where we ran antivirus checks on the manually copied over email files. Sneakernet. That's how it was. I doubt they get full access to their normal Earth-bound mailboxes. Astronauts don't have time to read much email while on a mission. I bet it is still highly filtered by a real person.

After I left, I understand they ran the JSC network into the entire control room for PCs so the controllers could check their email, but there was no mixing of flight control network with the JSC network (unless you wanted to be FIRED).

I suspect the LEO network has improved, but I have my doubts about non-LOS networking. There are always places in orbit with no ground contact on every orbit. Now, I'd probably setup an extremely low power linux server as an email gateway on both sides and use UUCP since the connection is "sometimes connected", not always connected.

Mail Server on both ends

[jfried]

mail server on the ground, mail server on the shuttle.

The mail queues up and you open up the connection between them certain times of day. Queue empties.
GZIP the link and your gold.

UUCP?

[shutton]

UUCP worked quite nicely in the days when links were ephemeral, slow, or generally unreliable. This seems like a lot of effort to solve a problem that existed 30 years ago, solved, and even adapted for RFC821 and its successors. There's a reason that Sendmail knows how to rewrite addresses!

Sounds like a bad idea to me

[Wannabe Code Monkey]

So, once a day they bundle a bunch of emails into a single .OST file and upload it to the shuttle. The astronauts then open that .OST file in their local copy of Outlook. And they have to shut down Outlook while the upload is in progress because of Outlook file locking.

In addition, communication with the ground isn't always possible (you'll hear warnings of LOS - Loss of Signal during mission communications) so standard methods of email transfer like POP/SMTP, IMAP etc might not be reliable.

If a 'Loss of Signal' can interrupt a POP session, wouldn't it also interrupt a file upload? Couldn't they just POP into the server on Earth once a day to grab their emails to be stored in a simple mbox or some such? Wouldn't this also eliminate the file locking issue as mboxes and Maildirs are pretty old and stable solutions that don't have this problem? This just sounds like someone wanted to use Microsoft Outlook no matter what and hacked together a procedure to use it even though there are way better approaches. And isn't the whole point of Outlook that it has a built in calendar and meeting request system and network folders? They're not even using those more advanced parts of it, they just need email.

Funny thing is, they got infected once

[Ilgaz]

In fact, you should be surprised that Windows is _STILL_ running after a Virus has hit the ISS orbiting the planet.

No kidding, Google it.

It is particularly sad that NASA IT guys aren't obviously that pathetic to license Outlook from MS. Something really going on there, a lot of open source software/operating systems has NASA contributed excellent code in them.

PS: I remember they also had Norton Utilities with "rescue diskettes" back in 1990s, it leaked while I was trying to find a way to manually uninstall norton...

Re:80's tech

[BitZtream]

They're doing it to save bandwidth. Yet they probably spend more on bandwidth dealing with human error issues in the process than they would if the system was engineered properly in the first place.

You don't see an issue because you aren't an engineer trying to save every drop of energy/bandwidth/processing time possible.

Basically, you're a java or C# developer when then need C and assembly developers with a clue.

Custom hacks when there are already systems (even build into EXCHANGE!) to do EXACTLY what they need to do are beyond stupid. Its one thing to use a custom hack so you don't get tied into a vendor, but their hack is entirely tied to their vendors so that rules that reason out.

Next you do it because you have a requirement that no existing solution fills in properly, which is certainly not the case here. As I already said, even Exchange will be happy to do store and forward batching on a schedule. A tiny exchange server (or a more efficient/less resource intensive alternative) on the space station could be designed to consume pretty much no energy unless it was actually in use.

In short, this is clearly something thrown together by engineers who knew nothing about the tools they were working with. Not their fault (probably), some douche bag manager probably didn't ask the IT guys.

The problem is, they went through effort and resources to make a system that is clearly less efficient than any of the possibly alternatives I can come up with.

Re:This hurt to read.

[LWATCDR]

"Why can't they just send a psk-31 HF radiogram? or the even more fault tolerant HF packet radio?"
The Ionosphere maybe. It will tend to block HF and one of the reasons HF is so good for long range on earth is that it can bounch and skip off the Ionosphere.
At best you would still just get line of site and the antenna would have to be very large.
You would be better off using VHF,UHF, or even Microwave since you could get more bandwidth.