In deepest darkest Surrey, an on-prem SAP system running 17-year-old software is about to die.... Surrey County Council is shopping around for a new £40m ERP platform as the existing SAP system has been running on in-house servers for so long it is in danger of falling over. A tender notice published last week says the southern England authority is looking for a "fully integrated ERP system on a software as a service basis …
Between 2012 and 2014, measures included migrating the system to the Surrey County Council data centre and bringing it under in-house operational control; creating a new procurement framework in conjunction with East Sussex County Council that was used by both Councils and resulted in annual cost saving of around 40 per cent for support and maintenance services, according to the council docs discussing the need for new support arrangements.In 2015, the council contracted with an independent support provider which effectively "froze" the council into its version of the software, with limited options for maintenance, upgrades or patching.
So bringing the system back in-house saved loads of cash so they decided to outsource it again in a way that prevents an upgrade...
A big victory for management there -->
Had something similar at one employer where a couple of companies who made software everyone liked were replaced. The new one was not as good and their software (designed to do both major functions of the predecessors) required propriety hardware too. I mentioned increased support costs, unreliability of previous versions, being locked into their ecosystem, loss of functionality etc. I was ignored because I was seen as being overly critical and didn't have as much clout as would be needed.to make a difference. What happened was interesting, after 6 months someone with clout said they wanted nothing more to do with it. There had been a few incidents and failures and the company over promising customisability or exaggerating existing features. At this point it was decided to keep one half of the suite for one function and get another supplier for the other. Impressive given the management speech about how wonderful and much cheaper it was to integrate everything under one supplier at the start of the project.
"So bringing the system back in-house saved loads of cash so they decided to outsource it again in a way that prevents an upgrade..."
I can see the logic. Sort of. Bringing it back saved money, ie reduced running costs. They now have a cost benchmark to work to so outsourcers bidding have to come under that benchmark making it it cheaper again. (naturally, it will cost more, but the headline "basic service cost" will be a headline grabbing saving. They;ll just try very hard to hide the costs of the extra little tweaks and services they forgot to put into the original tender which will increase, not decrease the annual bills.)
Management's view tends to be let's do the opposite of whatever the current situation is, so they can "create change" usually justified by "saving" money.
So if something is currently outsourced there are obviously gains to bring it in-house. If something is in-house - woah that's expensive - let's outsource it.
All the people in these positions do is try to justify their case using meaningless data that nobody can really be bothered going through with a fine tooth comb. Meanwhile collect a large salary, pension and other benefits. When everything goes tits up, resign, and leave other people to sort out the mess. Repeat every X years (where X depends on the circumstances / magnitude of what's being "changed").
"The existing server hardware has reached end of life and is on expensive extended support, with costs increasing every year."
Hmmm...I wonder what operating system has just reached end-of-life and is now on expensive extended support?
Surely they could have delayed this work for a few more years to ACTUALLY reach end of life/extended support/ok, this time we really won't support you any more...
LoL we had a parking system that was a bit like that but less important. All it did was count how many spaces were left in carparks then put that up on a sign.
It was running on an NT4 desktop. It hadn't been switched off for about 15yr and no one knew it was there since it just sat on a shelf doing its thing as teams were relocated around it. When it was found and needed upgraded to something current we couldn't since the company that wrote the software went out of business about 10yr before, there were no disks and no backup. No one knew if it was switched off if it would come back up again.
Problem was solved although I never did find out how. It might still be on that shelf yet.
Gluxo Smythe Klone had a similar issue where all the gates were controlled by two PC's (Not even in Security\Fire Dept offices). One day the inevitable power outage occurred & would not come back up allowing control of the gates.
It was out of scope for us in the IT Dept (Don't have anything to do with it, its not a Corporate issue said our boss wisely & we all concurred as nobody wanted to inherit that can of worms.)
We had a guy (Who claimed to be one of the minds behind IMDB), running parallel to us in IT, neither was he connected with Lab Support (IT) (I always got on well with Dr Derek & his team thus there was a little bit more help & assistance given in both directions than the powers that be would have liked if they had been made aware of our inter-departmental co-operation) who resolved the issue.
He discovered that one PC needed to get to the boot menu (NT4) before the other machine was powered on in order for handshaking to commence between the two as the second one came on-line.
Actually, SCC has offices in several places within Surrey - many of which should be considered deep and dark - as well as County Hall in Kingston which is, technically speaking , in Surrey - depending on which years' county boundary you use; as you say, That London annexed the land in 1965 because they were running out of green space so decided to steal someone else's. And with the current trend of trying to move everything out from the South, Surrey might get it back again when there's bugger all left in the centre of the Smoke and all the interlopers move back ..
Troll, 'cos there's still a few in the wilder areas of Surrey (rumor has there's a few still working the night shift at Ryka's...;-) )
Every time I read about how local and national government manage their IT, its seems their is a vicious cycle of short-term self-interested middle management, who seem follow a pattern:
1) new manager negotiates a fat bonus or promotion if he/she can sort out the IT budget
2) Manager engages external "consultants", who provide a flashy line-chart that proves savings through outsourcing
3) Manager proceed to outsource
4) Manager then proceeds to cheerlead about how brilliant it will be in the years to come (Whilst silently ejecting the in-house team)
5) Manager supplies evidence of savings (gained through redundancies... not the contract)
6) Manager collects bonus/promotion/senior role in other agency, and runs for the exit before the shit hits the fan
Now the fun begins...
7) Outsourcer starts increasing costs, due to "unforeseen issues" or "maintenance not covered by the contract"
8) Council/Agency realises it would be cheaper to in-house, but are stuck in a 5 year contract
9) Council/Agency take legal action or pay a golden-goodbye (to get out of the contract, X times its total value) the eventually in-houses
10) Council/Agency sets an unrealistic budget, provides no staff training, offers 50% market value salaries
11) Council/Agency realises their IT is in tatters, they recruit a hot-shot new manager, who has troubleshot a dozen different agencies.... in the past 5 years!
12) Goto 1...
Brand new, SAP-compatible hardware and the issue is done with. SAP doesn't run on special SAP PCs, now does it ?
Really, everything about SAP screams lock-in and expensive. On top of that, you have administrative muppets to manage the thing. They signed a contract in 2011 and nobody bothered to wonder about the hardware's end-of-life and what solutions were possible ?
This was not an unforeseeable problem. Somebody should have been paying attention, but no, back in 2011 it was Bonus Time, Yee-HAW !
Now you have paid millions for an ERP solution, and you're going to pay millions more for the same thing (but, in The Cloud !) because you couldn't be arsed to actually manage things.
Oh well, that how the market survives, with idiots paying the same thing several times over. On public money. No biggie.
The bigger issue is the software going out of support. For a payroll system in particular, there’s always a raft of new legislation changes that need to be catered for each year. Some of the calculations for deductions such as court orders are simply bizzare.
> They signed a contract in 2011 and nobody bothered to wonder about the hardware's end-of-life and what solutions were possible ?
To be fair, 2011 was right at the start of full-on austerity measures so there would have been no chance of spending more than the absolute minimum necessary.
I don't really recall a time in the too many years I worked there that BAU at SCC wasn't on max cost savings. Going away to conferences and things and finding out how much better funded one's oppos up north were was always kinda depressing. Even more depressing, though, was the way funds could always be found for exciting big projects that would look good on an execs CV, but never for little low profile stuff that might say reduce the number of helpdesk calls and give a better service to the users and thus taxpayers.
Is it really?
What's the best approach from this point on? Migrate the existing SAP to new hardware so that does not break, then upgrade the SAP software?
I don't know about SAP but I would expect them to issue new versions like stepping stones. You don't need every version update but you can only step so far. Yes the one they use may stop being supported in 2025 but the next one up from that would last longer.
The SaaS system would be good value...
Er, no. It doesn't matter how much bullshit you apply after this, still no.
Remember that "locked in with limited change options[1]" contract that dropped you in this pile of shit in the first place? Same thing.
[1] Unless you pay at "got you by the balls and we're going to squeeze really hard" rates.
<sarcasm/>
I thought the received wisdom these days was that the solution to a Government IT problem, any IT problem really, is Outsourcing, DevOps, and Cloud.
Repeat after me, in every meeting, at every opportunity: Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud. Outsourcing, DevOps, and Cloud.
Your training is now complete. Pick up your diploma at the printer on your way out.
That is indeed the received wisdom, shame everybody always forgets about security. And the problem with security is that it is very cheap if you embed it from the start and very expensive if you have to fit it in afterward. Been there and done that, the bill just about trebled.
I have only worked on one - the MOD CHOTS contract in 89-90. This was a wonderful initiative to get all three of our armed forces to use a common email and desktop infrastructure. I was the Uniplex (MS Office features/functions for UNIX on a Wyse 60) solutions expert and spent a happy 18 months working with Air Commodores, Air Vice Marshalls, Miss Moneypennys and lower ranks in pipe-smoke filled, oak lined offices at an MOD location in Central London.
The biggest issue in that time? We introduced automatic document labelling that put Classification/Caveat in the page footer which caused outrage as letters to bank managers and golf clubs had UNCLASSIFIED stamped on them.
I worked with an analyst who worked on that project.
He was sort of sad to see it complete as he was working on a large base west of london and if the officers wanted to discuss something they would get up, and take their dog for a walk to the other person's office.
He said the dogs were sad and real productivity fell.
Being forced to work with MS teams (say the name three times holding an image of the person you hate in your mind and an outlook reminder will pop up on their screen and remove any useful thought from their mind) things have not improved.
Does anyone else here remember when large payrolls needed a couple hundred staff with the skills to do it on paper? Now, you need the more staff, enterprise IT, licensing, and permanent upgrade cycles at higher TCO to do the same. ERP in general needs shooting and done well by a little indie developer. Can't be that hard...
Given the system is that old. The version they are on is only likely to run on certain OS/DB combinations. In order to get to a release that will run on a current OS/DB combination they will have to upgrade the software.
In fact they may have gotten to the point where there is no upgrade path at all.
Hardware is so old latest OS version will not run.
The OS they are on will not run on new hardware.
Current DB will not run on latest OS. New DB version will not run on old OS version.
The only upgrade path is to get old of hardware revisions that where released between old system and new.
To use cheaper hardware they may even have to do an "endian" migration.
If it's Oracle on HPUX they will be right Royally screwed.
In this case I'd say, definatly not an SAP problem, just them not keeping up with normal release cycles for a business critical system.
And now they will have to discover all of their business processes that were implemented in SAP, taking people out the business to help with the implementation as no doubt all of the original scoping work was sat on a shared drive somewhere, that has since been archived to /dev/null.
The business case warns the new system would potentially "reduce back-office staff costs" and "reduce the number of business support staff required to support a SaaS-based technology".
'potentially' is the key. This is standard boilerplate for proposals and means nothing. When did you ever see a government system that required less civil servants than the one that preceded it ? In fact it will take three times as many staff for the 5-year transition and twice as many thereafter.
Save your tax payers a pile of cash.
1. Clone the existing ECC6.0 servers into AWS. Hardware stability issues fixed and saved on hosting costs.
2. In 2025, when 95% of SAP customers are still not on HANA, gratefully accept the extension of support, even if they up the cost a bit.
Can I have the rest of the millions?
This post has been deleted by its author
Just wondering....is the SAP software HEAVILY CUSTOMISED?
Once upon a time, in a land far away, a huge retail organisation bought Peoplesoft and then had it HEAVILY MODFIED. For the next ten years, it was almost impossible to apply any patches issued by Peoplesoft. Duh!!
Then -- when the penny dropped -- said huge retail organisation paid a fortune to RE-IMPLEMENT the latest version of Peoplesoft with NO MODIFICATIONS.
Guess what.....huge consultancy bills at the beginning, during the ten year life of the modified sotware, and during the conversion back to "standard".
Why am I not surprised?
It is exactly the same with every software.
I worked at a company that was sold SAP as a solution to their problems. When the consultants came in (who may or may not have been the same as SCC) they asked 'what do you want', 'the same system on SAP' was the reply, and that was exactly what they got - an SAP system heavily customized to program in all the shit of the last system. 10 years later they stripped all of that out and decided to go with a fairly basic implementation fixing the underlying business processes and everyone was happy
"In 2004, the council worked with Capgemini to implement SAP R/3 to run finance, HR and payroll" see above.
So now the options are £2 million for a hardware upgrade, and ongoing software costs a bit more than before ..
OR £40 million for all the current buzzphrase gaga stuff that sure as hell won't match current work practices - just like the SAP dinosaur shit didn't.
Anyone noticed a strong smell of brown-paper envelopes?
It sounds like an interesting project to be honest. Taking an OLD old R/3 system and migrating it all out to SaaS is perfectly decent work and would grant a whole bunch of relevant contemporary experience and skills to all involved. If I wasn't already up to my eyeballs with an EBS migration programme I'd be tempted to give them a call...
There is no reason to use a common ERP system in the UK. None. Take Surrey for example.
It is unique.
Unique air, in fact as it passes over Surrey it's improved for the next borough
Unique Roads, the smell of freshly laid Surrey tarmac in the morning is like nothing else.
Unique Trees, we even have a hill named after a tree.
Unique Paths
Unique Bins
Unique People
Truly Unique and deserves its own stand alone system not some insanity like this https://www.theregister.co.uk/2012/08/02/oracle_r12_for_6_london_councils/
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
