Pentagon Develops F-35's 4th Generation Software

[maus92]

Block 4 is just a software update since the new hardware cam in Tech Refresh 2 (part of Block 3i).

Block 3F is scheduled to go IOC in 2018 and Blk4 is due for 2022/24 (parts A&B).

What's that ICP upgrade referring to in the Block 4 basics?

[cantaz]

We'll have to wait and see in four or five years, but because of the planned development time and the fact that they're starting now, I suspect that Block IV is a significant rewrite of the Block 3 code, and the programming team will ultimately have significant challenges to meet schedule, largely through the sheer number of existing functions that they'll have to duplicate in the rewrite.

That is literally wild, baseless speculation. All past issues where portions of a block needed fixing was explicitly done through patching of the problem version. There are incremental versions belonging to each block and sub-block that you seem completely oblivious to.

You are literally making sh*t up at this point. Not just that, you're not only accusing the JSF program of failing block 3F development, but then covering it up, before block 3F development is even complete. And you're accusing them of going to cover it up in a fairly stupid and attention-grabbing manner that completely antithetical to the way the software is arrangement.

What's that ICP upgrade referring to in the Block 4 basics?

The tech refresh that is to coincide with block 3I, or rather, the ICP upgrade originally scheduled for block 4 would've partly created block 3I when the schedule was revised. The ICP is designed to be plugged and play, to take advantage of the leaps in processing power. Therefore ICP upgrades are pegged to progress in the processor supply side. Block 4 software was pushed back, but ICP would've been able to stay in relatively the same place (i.e. to be done before 2015). Since the ICP upgrade stayed back, the port to the new ICP was included as part of 3I (also before 2015). According to the chart from Spud, it looks like the original block 4 ICP upgrade may have even been delayed due to 2B not being ready for porting to become 3I.

If there is to be another change to the ICP for block 4, the block 4 as it is now, it's not the same one on that link, but an entirely new addition representing processor technology of the 2020ish vintage.

[smsgtmac]

What's that ICP upgrade referring to in the Block 4 basics?

The F-35 is the first program that I am aware of that actually planned from the get-go to deal with the hardware upgrades (that always come) up front as planned "Technology Refresh" events

From http://thediplomat.com/2012/11/the-interview-lockheed-martin-talks-f-35/
Considering the cost of the aircraft, upgradeability is important as threats and threat environments can change throughout the planes lifespan. How upgrade friendly is the aircraft?

You are exactly correct. The threats that we face today are not the same threats that we will face tomorrow. Those threats are both improving and proliferating. The F-35 program has been designed with this reality in clear view. In the same way that the F-16V aircraft is vastly improved over previous F-16 aircraft, the Block 3 capabilities that represent the Initial Operational Capability will be expanded upon over the life of the F-35. These technology refresh upgrades are already planned for on the F-35 program, and will occur on a recurring basis throughout the life of the program. For participating nations, the opportunity to participate in these upgrades represents a tremendous benefit. On smaller programs, the proportional cost of the non-recurring engineering effort to integrate upgrades can be prohibitively expensive for any given customer. Comparatively, on a large program such as the F-35, the corresponding cost for a customer with a relatively small fleet will be significantly lower. Moreover, because the F-35 is the foundation for tactical air power for the U.S., the partner nations, and Israel and Japan going forward, should Korea select the F-35, they can be confident that the necessary upgrade investments will be made to ensure the continuing technological superiority of the F-35 throughout its life.

[smsgtmac]

Block 4 is scheduled already? Sigh. I was afraid they were going to re-write the F-35's code base... again.

Pure snark. Contributes nothing else.

While you are doing your remedial research on F-35 software development as penance, perhaps you will run across some clues as to why the software is complex, but is not insurmountable. Have you ever heard the acronym 'UML', and are you aware of 'visual tools' for C++?

Sometimes, I get the feeling people visualize legions of Code Gnomes laboriously hand-writing lines of code instead of moving icons around, drawing lines with arrows between them and hitting the 'compile' button. Simplistic explanation? Yes, But I think not nearly as simplistic as how most people visualize the work.

[spazsinbad]

I used to like watching the WHIZZs ruffle through absolutely yards and yards of machine code via a printer on a long screed of paper - to spot the one snag - then ask for more reams and reams of machine code printout (this was the IBM 360 era when 1 Mb of main memory was a huge upgrade, in a suitable large commercial refrigerator sized cabinet). What a life. Completely baffled me at the time but someone had to do it. Wot was worse were the punch cards. Chuck 'em in the air I reckon and then see what happens.

[maus92]

What's that ICP upgrade referring to in the Block 4 basics?

The tech refresh that is to coincide with block 3I, or rather, the ICP upgrade originally scheduled for block 4 would've partly created block 3I when the schedule was revised. The ICP is designed to be plugged and play, to take advantage of the leaps in processing power. Therefore ICP upgrades are pegged to progress in the processor supply side. Block 4 software was pushed back, but ICP would've been able to stay in relatively the same place (i.e. to be done before 2015). Since the ICP upgrade stayed back, the port to the new ICP was included as part of 3I (also before 2015). According to the chart from Spud, it looks like the original block 4 ICP upgrade may have even been delayed due to 2B not being ready for porting to become 3I.

If there is to be another change to the ICP for block 4, the block 4 as it is now, it's not the same one on that link, but an entirely new addition representing processor technology of the 2020ish vintage.

I think what you are saying is that the slide is old, and that TR-2 hardware arrived before Block 4 software was ready (and as it turns out, before Block 3 software was ready.) In other words, no new/upgraded ICP will be needed for Block 4 capabilities.

[SpudmanWP]

The ICP upgrade listed is specifically for comm updates and would not be part of a Tech Refresh (TR). If I had to guess, I would say it relates to the SATCOM system that is planned for Blk4.

Tech Refreshes have been par of the plan since the beginning. I was lucky to find a set of Audio/PDF presentations from back in 2008/2009 that went into the software and avionics plan. The plan was basically TR1 goes with Blk0.5 and TR2 goes with the last release of Blk2. Each TR covers two software blocks as TR1 goes with Blocks 1&2 and TR2 goes with Blocks 3&4. That has not slipped and is how it is happening.



You might have better luck getting the old presentations to show in your browser, but they will not in mine.

However, we did preserve the PDFs on the Program Docs thread.

viewtopic.php?p=161148#p161148

[cantaz]

The ICP upgrade listed is specifically for comm updates and would not be part of a Tech Refresh (TR). If I had to guess, I would say it relates to the SATCOM system that is planned for Blk4

The old software schedule does explicitly say that that the ICP update for the old block 4 is comms oriented, but the old block 4 was also to have completed development by 2015. Unless ICP development was itself behind schedule, to keep pace with processor development, the ICP update in old block 4 may have been merged with present day TR2.

My take is that, 1, the old block 4 ICP upgrade refers a board swap, that using the currently empty slots would've been characterized as ICP expansion, so the same activity as TR2. 2, ICP development is not completely pegged to software development, but to its own cycle. If hardware for old block 4 ICP upgrade was in everyway ready in time for TR2, which it would be based on original schedule, it makes the most sense to use it for TR2.

Ultimately, it feels a bit like reading tea leaves.

[SpudmanWP]

Hardware is VERY tied to software. This is evident by Block 3i being just a rehash (middleware adapted) of Block2B without any new features added. You cannot introduce new hardware without the software being ready to run on it.

Originally SATCOM was to be part of the SDD program thereby needing its associated ICP hardware as part of TR2. However, it was pushed to the right to Block4 and since the ICP upgrades related to SATCOM are not likely needed for Blk3, they decided to wait till Blk4 to install them (saving money).

[XanderCrews]

Sometimes, I get the feeling people visualize legions of Code Gnomes laboriously hand-writing lines of code instead of moving icons around, drawing lines with arrows between them and hitting the 'compile' button. Simplistic explanation? Yes, But I think not nearly as simplistic as how most people visualize the work.

So... So your saying its not like that?



@Rotoseq. you seem to be jumping to a lot of conclusions, bud.

[cantaz]

Hardware is VERY tied to software. This is evident by Block 3i being just a rehash (middleware adapted) of Block2B without any new features added. You cannot introduce new hardware without the software being ready to run on it.

Software is very tied to hardware. I'm not suggesting running 3i on the wrong ICP, I'm suggesting that comparing the original schedule for block 4 ICP upgrade vs current progress for TR2, the original ICP upgrade availability overlaps with the current TR2. I'm suggesting that it's feasible that the ICP used in TR2 was replaced with what was intended for block 4, with 3i then being accordingly ported to run on the newer ICP.

It's circumstantial, but if we consider the old schedule, the TR2 ICP was originally to be relatively short lived. TR2 ICP was originally to be superseded by block 4 ICP soon after end of SDD anyways.

[SpudmanWP]

The ICP is not a "single" card, but a cabinet with a whole bunch of different cards.

Here is the flow of early ICP upgrades showing the number of and different kinds of ICP cards.

[cantaz]

I just found that post of yours as well. I said "a board swap" to characterize the overall activity, not the number of boards involved. You said that blocks 3 and 4 resides on TR2. Specific to the notion that moving any block 4 ICP work into TR2 would've been less cost effective, I disagree. If ICP development was on track based on original schedule, it made more sense to adapt the block 4 ICP upgrade during TR2, port 2B to that, so no ICP change is required until block 5+. Excess/early processing capability become a hedge against unexpected software demand in 3F, and won't be excess for long given the relatively short time until block 4 software.

Given that no announcement has been made for any changes to ICP post-TR2 other than what was indicated in the very outdated software schedule, well, we have to wait and see.

[maus92]

The LRIP VI jets coming off the line shortly will have TR2 hardware, and cannot be accepted without a certified version of 3i, I suppose for contractual reasons. The Gilmore report noted that initial attempts to load an early version of 3i failed (on the F-35C mission systems jet modified with TR2 hardware,) but nothing much else has been said, other than 2B software won't delay F-35B IOC, presumably using early LRIP (<VI) TR1 aircraft.

[smsgtmac]

I used to like watching the WHIZZs ruffle through absolutely yards and yards of machine code via a printer on a long screed of paper - to spot the one snag - then ask for more reams and reams of machine code printout (this was the IBM 360 era when 1 Mb of main memory was a huge upgrade, in a suitable large commercial refrigerator sized cabinet). What a life. Completely baffled me at the time but someone had to do it. Wot was worse were the punch cards. Chuck 'em in the air I reckon and then see what happens.

A little off topic, but it highlights how fast things change computing-wise and I know what you mean about the 'cards'. Guy I worked with for about 5-6 years got into the modeling game in the late 60s. He was quite a character, Annapolis Pistol team, and was a former Navy wargame wonk who specialized in being the OpFor. He had 2 cruises off the coast and an exchange tour with Army Intel in-country during Vietnam. I learned a lot about Operations Analysis from that guy. At the time he retired with my company he was the absolute reigning world's greatest authority on the use and abuse of the DoD LCOM model and had a permed Afro (looked like Bob Ross) and was a happy guy. All the DoD guys who originally developed LCOM were long gone, and he had worked with LCOM 'forever', transitioning from the mainframe days to the early server era to the PC, but he always had a bunch of blank punch cards in his shirt pocket he used as notecards. He had saved crates and boxes of them from decades before instead of throwing them out and loved to talk about how wild it was to do a model run. It was just as you described. Even on the PC, he still referred to changes to any model as 'change cards'. When he first got the model on a PC, he'd start the 'run' on a Friday and hope there was no program crash or power outage over the weekend. About 2000 I got him one of the 'new' (to us anyway) dual processor PCs and I was his friend for life. First time it was EVER possible for him to do a model run in under a half-day: before it was the 'cards' slowing him down and then it was the lack of processing power.