Chair of Met Office board

I doubt that there will be many applicants from YBW forum members, but, obviously cannot rule it out. The application details shine a light on the Office that might be of interest, Role details – Chair of Met Office – Apply for a public appointment – GOV.UK. Interesting that they are going to use Microsoft cloud computing.
I wonder if it worth bunging in an application, as it is just down the road.

Not really a big choice for cloud computing, Microsoft, Google and Amazon. All of whom got a tad upset when I pointed out to them that cloud computing is 'modern' name for a mainframe. Then stood there open mouthed when I explain to them what we were doing 50 years ago!
 
I wonder if it worth bunging in an application, as it is just down the road.

Not really a big choice for cloud computing, Microsoft, Google and Amazon. All of whom got a tad upset when I pointed out to them that cloud computing is 'modern' name for a mainframe. Then stood there open mouthed when I explain to them what we were doing 50 years ago!

Modern cloud computing is very different to monolithic mainframes, I also worked on a mainframe and now work with serverless, scalable, on demand systems. The architectures are very different.
If you are talking about lift & shift, then I would agree with you, it's the same, but building real cloud serverless applications is very different.

As for them choosing M$, it's a result of using big outsourced companies such as Logica that are so ingrained in M$, and most of the time still build monolithic lift & shift systems. Basically reinventing the wheel.
 
Modern cloud computing is very different to monolithic mainframes, I also worked on a mainframe and now work with serverless, scalable, on demand systems. The architectures are very different.
If you are talking about lift & shift, then I would agree with you, it's the same, but building real cloud serverless applications is very different.

As for them choosing M$, it's a result of using big outsourced companies such as Logica that are so ingrained in M$, and most of the time still build monolithic lift & shift systems. Basically reinventing the wheel.
In operational weather forecasting, there is a lot of lift and shift, sequential computing to provide data analyses for a mass of parallel processing.
 
In operational weather forecasting, there is a lot of lift and shift, sequential computing to provide data analyses for a mass of parallel processing.

Understood, I'm just surprised and not really surprised that they are not taking advantage of the big data processing and analytics that AWS has developed. It's leaps and bounds ahead of M$ and Google.
 
Understood, I'm just surprised and not really surprised that they are not taking advantage of the big data processing and analytics that AWS has developed. It's leaps and bounds ahead of M$ and Google.
I am out of my depth when it comes to hardware systems although I see that ECMWF has gone down a different route. But, of course, they are different beasts with different aims and responsibilities.
 
I doubt that there will be many applicants from YBW forum members, but, obviously cannot rule it out. The application details shine a light on the Office that might be of interest, Role details – Chair of Met Office – Apply for a public appointment – GOV.UK. Interesting that they are going to use Microsoft cloud computing.
My understanding is that the computer the met office is using is a Cray XC40 super computer.
Not really your average cloud computing platform.

The Cray XC40 supercomputing system
 
My understanding is that the computer the met office is using is a Cray XC40 super computer.
Not really your average cloud computing platform.

The Cray XC40 supercomputing system

The XC40 is an impressive bit of kit, but nearly 8 years old now.

This is from September 2021:
Running the Harmonie numerical weather prediction model on AWS | Amazon Web Services

TLDR:
Comparing across the systems AWS c5n.18xlarge (Intel Skylake based) has a 43% performance advantage over the Cray XC40 (Intel Broadwell based) at 145 instances (nodes) and is 16% better over Cray XC50 (Intel Skylake based) at 86 instances (nodes).
 
I’m pretty sure they already use Microsoft, as did ECMWF. Of the three clouds it’s probably the better option right now for various reasons but the investment in GPU compute is astronomical including a £1Bn data centre in “London” being built right now to house mostly AI workloads on GPUs.

We also do have Cray as an option in Azure for what it’s worth.
 
Top