OVH Community, your new community space.

FMECA (Failure Mode, Effects, and Criticality Analysis)


Myatu
29-12-2009, 18:43
Quote Originally Posted by oles@ovh.net
"ovh please don't propose any innovation, just the old servers"
Je ne vous ai pas dits pour cesser d'innover. En moins d'une année vous avez apporté beaucoup de modifications à votre gamme de produits et ceci le rend difficile de faire le plan d'action à long terme avec OVH. Par exemple, je prévois de commander 10 serveurs le mois prochain et de les compter payer «X», puis soudainement vous dis qu'en 3 jours ces serveurs aura différentes caractéristiques et le coût «X+Y» et ce qui était standard maintenant coûte un «Z» supplémentaire. Par conséquent ma référence à construire une maison sur le sable mouvant -- pour employer OVH comme base aux plans à long terme est comme builiding sur le sable mouvant. N'avez-vous pas voulu réduire le chiffre d'affaires élevé ?

Je peux comprendre que vous devez vous déplacer avec des périodes. Je ne peux pas m'attendre à ce que vous utilisiez toujours 8086 ordinateurs avec le 1 réseau coaxial de Mbps pour des sites Web. Mais au lieu de le changer tellement souvent (et sur une telle brève information) ne pourriez-vous pas faire le ce graduellement et en conformité avec des plans existants ?

--- Oops, forgot the original English version again: ---

I didn't tell you to stop innovating, but to sit still for a moment. In less than a year you have made many changes to your product range and this makes it difficult to make long-term business plan with OVH. For example, I plan to order 10 servers next month and expect to pay "X", then suddenly you say that in 3 days time those servers will have different specifications and cost "X+Y" and what used to be standard now costs an extra "Z". Hence my reference to building a house on quicksand -- to use OVH as a foundation to long term plans is like builiding on top of quicksand. Did you not want to reduce the high turnover?

I can understand that you need to move with times. I can't expect you to still use 8086 computers with coaxial 1 Mbps network for websites. But instead of changing it so often (and on such a short notice) couldn't you do this gradually and in line with existing plans?

slayer2005
29-12-2009, 16:10
Would like no setup fees or at least just no setup fees for the low end servers.
And the abbility to pay via paypal.

RapidSpeeds
29-12-2009, 11:27
Quote Originally Posted by oles@ovh.net
"ovh please don't propose any innovation, just the old servers"
You are missing the point... I think what he is simply trying to say (or I am anyway), you launched these HG Servers for businesses 2months ago, yet now you are saying you want to offer more customizable servers for them? Yet, us 'little' guys who helped build up your company in the first place are being hit with set-up fee's and options that are here one minute, gone the next. (old kimsufi plans/usb hdd etc)

Why don't you allow users to upgrade to 1Gbps switchport for example (instead of *having* to be placed at order, or face a set-up fee for exact same server with 1 extra cable), that would be a start..

Or how about OVH buy extra IPs, and allow customers to purchase them direct from OVH instead of RIPE, for a standard non-rip-off fee?

I could go on and on... but I hope these are the kind of 'customizable' options you are thinking about...


----------


Vous êtes hors sujet ... Je crois que ce qu'il essaie tout simplement de dire (ou je suis de toute façon), vous avez lancé ces serveurs HG pour les entreprises il ya 2 mois, mais maintenant vous dites que vous voulez offrir des serveurs plus personnalisable pour eux? Pourtant, nous «petit» gars qui ont contribué à bâtir votre entreprise en premier lieu, ils sont frappés avec set-up frais et les options qui sont ici une minute, allé l'autre. (plans Kimsufi ancienne / USB HDD etc)

Pourquoi ne pas vous permettez aux utilisateurs de passer à switchport 1Gbps par exemple (au lieu de devoir être mis à l'ordre, ou face à une taxe d'installation pour le serveur exactement la même avec 1 câble supplémentaire), ce serait un début ..

Ou que diriez-OVH acheter IPs supplémentaire, et permettre aux clients de les acheter en direct chez OVH au lieu de RIPE, pour un type de non-rip-off des frais?

Je pourrais continuer encore et encore ... mais j'espère que ce sont le genre de 'personnalisables' options que vous pensez ...

oles@ovh.net
29-12-2009, 11:12
> Cessez de changer votre gamme et prix de produits tellement souvent. Si
> vous voulez vraiment être plus de B2B amical, alors j'ai besoin de plus
> de stabilité sur votre extrémité. C'est pourquoi je n'ai jamais commandé
> plus d'OVH, parce que la semaine prochaine il n'existe pas ou il coûte
> plus dû pour installer des honoraires.


alias

"ovh please don't propose any innovation, just the old servers"


RapidSpeeds
29-12-2009, 11:10
Vous ne construisez pas une maison sur le sable mouvant. <-- lmao

very accurate.. Ovh was built on some basic principles and now they are shaking those very foundations

Rahxie
28-12-2009, 22:28
This sounds interesting, I'm loving how OVH are adding more and more services lately.

Myatu
28-12-2009, 22:23
Ce genre d'information sont normales dans les relations B2B car il y a une totale transparence et les contrats qui permettent savoir à quoi s'en tenir.

...

le truc est là et vous le savez, pas besoins de chercher les explications en minuscules en bas de page
Est-ce qu'ainsi pourquoi c'exactement l'opposé est à OVH?!?

Cessez de changer votre gamme et prix de produits tellement souvent. Si vous voulez vraiment être plus de B2B amical, alors j'ai besoin de plus de stabilité sur votre extrémité. C'est pourquoi je n'ai jamais commandé plus d'OVH, parce que la semaine prochaine il n'existe pas ou il coûte plus dû pour installer des honoraires.

Vous ne construisez pas une maison sur le sable mouvant.

J'ai oublié : Nouvelle année heureuse !

Andy
28-12-2009, 21:57
I think what should happen is drop the setup cost for "pre-specced" servers, e.g. ones you have determined are good setups, and add a setup fee for those who want their server custom specs. That way you can decide what CPU you want, what amount of memory, what size and how many hard drives you want. That sort of setup would be perfect for me.

My perfect server for my 'current' needs would be:

Core2Quad Q6600
4GB RAM
1x80GB SATAII (boot)
4x1.5TB SATAII in 2xRAID0 arrays with manual syncing
100Mbps with 10TB B/W

That's easily do-able for not much cost than my current server.

If you're worried about a customer not staying for long, then tie them into a 6 or 12 month contract on those custom servers, paid monthly. In all fairness only a fraction of people can pay for 6 or 12 months up front.

yonatan
28-12-2009, 21:33
The problem with dedicated servers is that there is no " one size fits all " , we can clearly see that, and sometimes feel that.

I had a project which was lost for OVH due to the fact that the client needed an MG server but with more disk space ( 2X1500gb did not fit ) , and the next option was the HG which the setup fee drove him away from this investment.

when we talk about Large databases and a need for storage on the same box, we cannot find that in OVH.
currently i got this client on my rack over at my local ISP (he got 2 machines one for database one for storage with more then 1500gb ) , but i really wanted to push him to work with OVH since the costs are much lower, and the quality is great.

I would like to see a hybrid server a mix of the SP storage combined with the EG-ssd for example, 80X2 1500X4 and the new xeon - that would fit exactly!

as for shared hosting, the same, why not a hybrid machine? ssd+sata

/var /usr /tmp /opt will fit the ssd small size, while /home must be a large place for the content - i see this as the perfect server for shared hosting.

EG/MG Bestof + 2X80gb intel SSD

maybe create a subversion for each range? ( as a fix setup not configurable )
SP hosting
EG hosting
EG Streaming
EG - DB
MG hosting
MG Streaming
MG - DB

etc . . . there are many types of setups for each type of client.

the problem i see here ( form the side of OVH ) will be the setup efforts.
add parts / change parts / remove parts , from the live racks. - you are talking about 70,000 servers and counting , i think you are aware of the manpower needs and clients do not like to wait too long.

so you have your base ranges, but who will manage all the hardware work? we already feel how you guys work hard there to get the stock versions of your servers out, so add more features? - that equals man power, which equals extra fee for the clients, and in general equals MORE BACKLOG! we don't want\like that...

you talk about FMEA, but we don't have the details from your side, what is the current situation, i get my fair share of requests from end clients for products and solutions, the current way is to lead the client hand by hand and tell him what he needs from our inventory , and not letting him look at anything else, the major risk here is the price getting un-attractive if you try to play with the stock plans

This is not a new born baby, OVH is a living creature, keep that in mind.
the setup fee is one example for a killer, its a major risk but the abuse factor is more powerful so for the kimsufi range we can understand it, but why MG/EG need to pay the price also?


the RPS example is great, its a highly technical solutions , but is provided in a price for novice users who don't know what it is.

iSCSI is not a common word such as "intel / amd / DDR2 / 100Mbit" , that's the main point of failure , people look at the great CPU of the RPS3 for example, take it and when the iSCSI fail , what do you tell them? what can you do about it?

well i think my point is clear, if you are not sure where to sail the boat, keep it in the port.

Andy
28-12-2009, 21:19
Quote Originally Posted by rapidspeeds
looks like it - except it won't be custom to our needs, it will be what ovh thinks we want.

Just like he thinks we wanted set-up fee's
+1 :d

rickyday
28-12-2009, 19:35
No Set Up Fees!

Merry Christmas and a Happy New Year!

RapidSpeeds
28-12-2009, 19:31
Looks like it - except it won't be custom to our needs, it will be what ovh thinks we want.

Just like he thinks we wanted set-up fee's

Razakel
28-12-2009, 18:10
So you're planning on introducing server customisability?

sic
28-12-2009, 17:49
I am not sure but i think i might have a good idea for you oles?

NO SET UP FEES!

Mac
28-12-2009, 17:30
Blank Post :P

RapidSpeeds
28-12-2009, 17:25
Quote Originally Posted by MicroChip123
My suggestion is to have no setup fees for people who already have servers.

Block port 25 unless they provide ID.

And no setup fees on RPS
No set-up fee's (at the very least reduced) for people who already have servers (reduced fee's for us resellers when the program is finally up and running)

Block port 25 unless they provide ID and/or purchase the OVH Email Gateway package.

The RPS have always been considered 'play' servers - so to have a set-up fee is ludacris... even a simple £10 set-up fee would suffice, and still generate more revenue to line your fat pockets octave..

Nice and simple

Mac
28-12-2009, 17:23
Yes Oles seriously....Please do something about the Setup fees..
For customers suppose say with Above 50 Servers or so...remove the setup fees or so....
You have Bigger brains than us...lol
I am sure you will come out with some Idea....

MicroChip123
28-12-2009, 17:13
Hello,
We are conducting an internal debate that is now seeking to expand our
customers to have your opinion on the issue.

Ovh offers product lines that are fairly Linear (domain, mutu,
rps, dedicated housing) but also range Linear SP, EG, MG, HG. We
quickly understands that dedicated is better than the RPS and that the MG is better
qu'EG.

We want to create new products that are not Linear. They will
be there to respond to specific needs. For example, for memcache
there is no need of CPU or disk, but lots of RAM. For
storage backup there needs a lot of drive and performance
are not so important, while in production storage is
Conversely.

We arrived at the idea of creating a "standard" performance servers
being proposed. Something equivalent to "ABCDE" we can
see in the trade. The purpose of highlighting the performance
of certain components we use in certain ranges and meet
some kind of answers
- "Why the server is less expensive than this one"
- "Why that server is more expensive than this one"
- "Why is better?"
- "What is worse?".

For example, the performance of disks between
- SATA2 5400RPM
- SATA2 7200 rpm
- SAS 7200 RPM
- SAS 15000tr/min
- E SATA2 Intel SSD 80Go/160Go
- Intel SSD SATA2 M 32Go/64Go
are not the same. And the same goes for iSCSI / FCoE and NFS / CIFS.
But also on the CPU, network, firewall etc.. View mutu ...

The performance is not bad, but they continued to think and we said
to include in this reflection "risk". Basically, give information
on risk, probability of failure and encrypt the information noted. So,
each client may use this information to make the best choices
qu'Ovh proposes solutions based on its own imperative. We believe use
methods derived directly from the car industry and railway called
FMEA (Analysis of Failure Mode, Effects and Criticality)
English equivalent: FMECA (Failure Mode, Effects and Criticality Analysis).
More:
http://fr.wikipedia.org/wiki/Analyse...leur_criticité
http://en.wikipedia.org/wiki/Failure...ality_Analysis

We believe that such information would allow for better information on this qu'Ovh
proposes exactly. Already this would adjust your budget to your needs. Avoid
purchase of unnecessary parameters that you do not need. And in contrast to
budget on other criteria that are important to you. Such information
are normal in B2B relationships because there is total transparency and contracts
which allow to know what to expect. And we have more and more customers who
ask us to sign commitments. Also in reading your feedback, it was not bad
new ideas about new products. The problem is that in a range of
"non linear" which is specialized servers for patches, there may have
see the confusion of distrust on the "cheap", the "dear", the "cheaper" and
"dearest". And therefore we need more transparency. If tomorrow you see a server with
24TB of disk for 30euro, you will say "where is the trick?". That is what
to avoid bringing forth the answers, the stuff is there and you know, not
needs to seek explanations in the tiny bottom of page ...

What do you think? If you have any experience with FMEA and what you want
have, please contact me also (oles@ovh.net).

Regards
Octave
My suggestion is to have no setup fees for people who already have servers.

Block port 25 unless they provide ID.

And no setup fees on RPS

oles@ovh.net
28-12-2009, 17:06
Hello,
We are conducting an internal debate that is now expanding to find out the opinions of our customers on this issue.

Ovh offers product lines that are fairly Linear (domains, hosting, rps, dedicated hosting) but also a linear range SP, EG, MG, HG. We
quickly understand that dedicated is better than RPS and that MG is better than EG.

We want to create new products that are not Linear. They will be there to respond to specific needs. For example, for memcache
there is no need of CPU or disk, but lots of RAM. For storage backup alot of disk space is needed while performance is not so important, while in production storage it is the inverse.

We came to the idea of creating "standard" performance servers. Something equivalent to "A B C D E" which can be seen in trade. The purpose is to highlight the performance of certain components that are used in certain ranges and respond well to certain questions:
- "Why the server is less expensive than this one"
- "Why that server is more expensive than this one"
- "Why is it better?"
- "What is it worse?".

For example, the performance of disks between
- SATA2 5400RPM
- SATA2 7200 rpm
- SAS 7200 RPM
- SAS 15000tr/min
- E SATA2 Intel SSD 80Go/160Go
- Intel SSD SATA2 M 32Go/64Go
are not the same. And the same goes for iSCSI / FCoE and NFS / CIFS.
But also on the CPU, network, firewall etc.. See hosting ...

The performance is already not bad, but we continued to think and as we said we want to include in this reflection "risk". Basically, give information on the risks, probability of failure and known encryption information. So, each client may use this information to make the best choices, from the solutions that OVH proposes, based on their own mandate. We believe in the methods derived directly from the car and railway industry called FMECA (Failure Mode, Effects and Criticality Analysis).
More:
http://en.wikipedia.org/wiki/FMECA

We believe that such information would allow for better information than what Ovh proposes exactly. Already this would adjust your budget to your needs. Avoiding purchase of unnecessary additions that you do not need. And in contrast allow you to use your budget on other things that are important to you. Such information are normal in B2B relationships because there is total transparency of contracts which lets you know exactlywhat to expect. And we have more and more customers who
ask us to sign commitments. When reading your feedback, it was not bad regarding new ideas about new products. The problem is that in a "non linear" range which is specialises in task servers, it can be confusing to see or distrust the "cheap", the "dear", the "cheapest" and the "dearest". And therefore we need more transparency. If tomorrow you see a server with 24TB of disk for 30euro, you will say "where is the trick?". That is what we want to avoid by bringing forth the answers, the solutions are there and you will not need to seek tiny explanations on the bottom of page ...

What do you think? If you have any experience with FMEA and if you want, please contact me also (oles@ovh.net).

All the best

Octave