We are in the process of migrating this forum. A new space will be available soon. We are sorry for the inconvenience.

Delivery of servers 2009 2010


oles@ovh.net
12-20-2009, 05:40 PM
Hello,

Since the establishment of installation costs, we are finally beginning to see the end of the tunnel in shipments of servers in 1 hour. Despite the fact that 92% of orders are actually carried out in less than 1 hour, since the month of September 2009 we had way too regular long delays on deliveries of some servers. Currently, we still have 50 servers to be delivered late, which is nothing compared to some moments ... but still these 50 servers and every day, despite the fact that we book every day, there's always the 30-70 servers ... Some references and only a few hours/days, but it's still there ...

Why do we have (had) these delays? With hindsight, we believe that the crisis has had a pernicious effect on all manufacturers of computer parts (the whole global industry?). One had the impression that the production plants worldwide has been clamped on "sales forecast this month the worst." Normally there are ups and downs. Once managers have anticipated that there will be that low. For example, factories could only run at 100% and have the best possible financial return (that dream factory working 100% of the time?). It is as if capitalism has said that the investment "is too risky" (!) While the interest rate has never been "zero". It seems that many manufacturers have preferred to sacrifice profitability for growth but also sell the parts immediately without any period of stock instead of taking risks to produce parts that put a little time to sell. I have seen this in my childhood, it's called communism:
- hello, I would like 2 loafs.
- There is none.
- 1 loaf then.
- There is none.
- Then what is there?
- There's me

... a Polish joke. Pass

This is why the market is empty of coins. There is now some equipment ordered 3-6 months in advance. This greatly complicates our internal management, this has removed a part of our flexibility, but mostly we had a terrible time producing as many servers as you need. We were out on our feet, because of pressure, pledged on firm numbers, prepaid orders, and I spent the best ... There was no loss of life and yet ... And we have greatly limited the cases, not perfect, but by as much as we can, while being abused by customers. "I like the crisis.com" ...

At the same time, the crisis has "slipped" our customers down, meaning that the client was looking for a "cheap" server, the customer can not pay at all. He has more money. He is now seeking a free solution, it can not pay at all. At the same time the companies who did not want to work there to rent (and sought only by the bay, the true, the server manufacturers, real, heavy investment and long) have become clients of big servers. The consequence is that we now produce many large servers and small servers, cheap, nobody wants it. Apart from spammers, hackers and those with transient activity, nobody wants these servers. Hence installation costs which, after 1 month of life, giving excellent results. Better than our wildest dreams ... In short, we keep installation costs.

How will it happen in 2010? It is not known yet how the supply of parts will be in 2010. Apparently it will not be better. This means we will have to continue to do miracles to provide sufficient room (even VoIP phones we had to buy the stock of parts for 12 months of sales ...). At the same time, we know that until September 2010, we have only the RBX3 datacentre that we can do with the new servers i.e. between 20,000-30,000 new servers (depending on type of servers, 20000 EG / MG or 30000 SP) over 8 months. There are 2500-3750 new servers each month. With installation costs, we went from 5300 servers shipped in November to 4100-4400 provided for in December ... and therefore it was a few months to find miracles, creative solutions in order to make RBX3 > 35,000 servers. All this knowing that OVH now houses 70,000 servers and therefore this amounts to 50% growth in 8 months. We talk about incredibly important figures that we require to manage OVH as a Formula 1 launched at 400km/h on mountain roads in winter on the ice. For now, it is expected to succeed at the new challenges that 2010 seems to throw us. But please, if you need a server in the coming months, consider providing the command with a few days / 1 weeks in advance. Indeed, within a few days we will launch the Vrack (Virtual Rack, Virtual Bay) in direct control: a control system that creates a cluster of several floors full of servers (Web, SQL, STORAGE, BACKUP) with distributions of charge, firewalls, SAN / NAS Storage etc. The result: 1 control, 3 clicks and it's up to 120 servers to be delivered preconfigured and Vrack in 1 hour to 1 single customer ...

.... we will laugh in 2010 ... long as it lasts! we love it ...

All the best,
Octave

oles@ovh.net