Hi,We are wondering what kind of rates and/or contracts people have with server/platform management for large scale ecommerce sites? Most of the server management discussed and offered on this forum is for small tasks, such as single servers, for installing software, upgrade kernels, setting up apache etc, ie: for people who have no technical know how when it comes to managing a server.However what kind of industry rates would one be looking at to manage a complete platform which consists of clusters of servers in multiple data centers spread across the world, with custom replication systems, local and geographic failover etc. All of this runs a single website which generates over $10+ million per year in sales.How does one go about working out a contract rate for doing the complete support (maintenance, development, and monitoring) of such a system? Is it typical in industry that such responsibility should be determined or calculated by a percentage of the sales and a rate paid monthly?Look forward to your opinion and thoughts.Thanks,Paulwell personally i think you should be visiting the offices of any server management company and discuss that with them.a web site with $10+ million per year in sales shouldn't choose a server management company from a forum as they usually can't take the risk of finding out that the company really sucki have only one server and i had a very bad experience with a company that most of the ppl told me they are very good and professionalso i guess its better than going and checking your self rather than waiting for the opinion of some ppl that might turn to be wrongthats my 2 cents 
a web site with $10+ million per year in sales shouldn't choose a server management company from a forum as they usually can't take the risk of finding out that the company really suck
There's nothing wrong with getting opinions on this here. Really, this is the best place FOR that kind of thing.
so i guess its better than going and checking your self rather than waiting for the opinion of some ppl that might turn to be wrong
There, you're truly correct. However, there are ways to verify this and check that the company isn't lying on their own webpage. NEVER trust anything that a company puts out on their own page, unless they offer alternative URLs for verification as well.
Now, back to the topic @ hand:
Most of the server management discussed and offered on this forum is for small tasks, such as single servers, for installing software, upgrade kernels, setting up apache etc, ie: for people who have no technical know how when it comes to managing a server.
This may be true for MOST, but I can think of a number of them that actually offer more than just that. You'll want to stay away from the 'cookie cutter' companies though.
A suggestion:
Go through the server management companies listed here (<!-- m --><a class="postlink" href="http://www.webhostingtalk.com/wiki/Companies_providing_server_management">http://www.webhostingtalk.com/wiki/Comp ... management</a><!-- m -->). Find the ones with websites listed and view them.
Get a list of the ones you liked and send them a mail (best bet is to use a 'contact form' if they have one). Give them the situation, and ask them for a price quote.
Even better? Ask for a call back to discover the potential relationship with your company.
Pick one, give them one (just one) of your servers to manage for a brief period (say 30 days). This way you will know if you picked the right one, and how well you work with the company, and they fit your needs.
As for cost? The response will really vary with what you need. Any company that quotes you a specific 'cost' without getting to know your needs personally is only going to cause more problems in the longrun.Greetings Paul:
Depending on the services you need, a management team dealing with mission critical servers and services, offering 15-minute or less response time 24x7x365 et all can run anywhere from $100 to $250 per hour.
Thank you.Agreed that there is nothing wrong with getting information or even hiring support/engineers from here as many people here are private contractors, the same sort of people you find at IBM, EDS, etc. I know for a fact many people here are former employees of those companies..Just perform your due diligence as you would with anything business related. The single most important thing for your project are the people that work on it, so make sure you can -clearly- communicate your needs and expectations prior to speaking with an engineer.Kind Regards,Thanks for the feedback so far, we have also received a number of good PM's. If you are a Server Management provider who has experience with providing management and 24/7/365 monitoring for clusters (preferably geo load balanced), then feel free to send us a PM and we can talk further about our situation in detail.Rates will vary greatly for something like this. A 3-4x range would not be surprising. Also, you are talking about several different services: DB admin, sys admin, network admin. Each group will have specific rate ranges.
I would expect for an operation such as this you would have an internal sysadmin to co-ordinate this effort (perhaps that you?). I don't want to offend by stating the obvious, but you would be surprised how many requests we get at rackAID that simply make no sense from a technical standpoint or are filled with mis-information.
I would also look for a firm that can list specific personnel that will typically handle certain issues.
Also, your SLA's response time requirements will highly impact your costs. There is a huge cost difference between having someone on-call vs. on-demand.
Don't buy more than you need. Some all-inclusive plan may sound great but can often mean you are paying for more than you need.
You may also want to search around for risk calculators. How much downtime can you afford? I wrote a blog post about this some time ago:
<!-- m --><a class="postlink" href="http://www.rackaid.com/resources/rackaid-blog/industry-insights/hosting_costs_getting_it_right/">http://www.rackaid.com/resources/rackai ... _it_right/</a><!-- m -->
The post is geared to smaller operations but some points are the same.
I would say the idea of downtime even extends to your outsourced IT group. What if they go "down" and cannot resolve a problem for you. What are your fall back options?
Lastly, don't overlook your product vendors. If you bought GSLB, firewalls, etc. Then your vendors may have some support plans or recommended providers.
So, in short ... I don't think there is some ball park number. The upper bound of what you can spend should be set by business objectives not by IT objectives. I would put together a plan, put it out there and see what it costs. Then see if this meets your budget. If not, then some of your business criteria may need to be re-adjusted.A site with $10+ Million in sales is all relative. What kind of transaction volume are you actually doing? Are these small average ticket sales or big average ticket sales? $50 average ticket would be less than 600 transactions per day, whereas $25 average ticket would be over 1,000 transactions per day. Is your volume consistent day to day or do you see large spikes in sales during certain periods such as major holidays?How many total servers are your cluster? How many data centers are involved?I wouldn't say that there is a standard rule of thumb in terms of percentage of sales toward development, management and maintenance costs as every organization is different. What you need to weigh are the costs of bringing such expertise in house versus outsourcing it. As others have suggested, you may have a mix of outside support personnel in addition to on staff personnel. Depending on the variables above, you could be allocating anywhere from 2% to 5% of your revenues towards Web site continuity which would translate into $200K to $500K/year. On the flip side, we have customers with multimillion dollar yearly revenue e-commerce sites spending less than 1% of gross sales in terms of maintenance, support, updates, etc.Do you have any idea what you are spending now?I think its far more typical for a site making $10m+ to have everything but hands/eyes support in-house, relying on the local datacenter providers for hands and eyes, and possibly network management alone.. everything else being handled by staff that work for you, and thus have the site as their one and only priority.$10m in sales would warrant hiring your own server management team so you can be 100% sure of what's going on.
If you don't have that team in-house, then you would have it with whatever outsource provider you hire.
What you might want to do is have a chat with a reputable and established company in the industry that specializes in this sort of thing, such as <!-- w --><a class="postlink" href="http://www.rackspace.com">www.rackspace.com</a><!-- w -->.
They can set you up with a dedicated team of server admins to monitor your application and servers 24x7. It will not be cheap.Or do both of what mrzippy suggested... well sorta.I've finally found the balance of both choices for my company (for our current needs) with RackSpace. We use RackSpace's managed hosting for 2 of our 3 global clusters (EU + NA) and have 1 full time Sr Sys Admin on staff to interact with the RS techs. This allows a company to have full control and faster reaction on most things but also have constant monitoring and 24x7x365 hands on the server level.It's paid off numerous times already where RS have caught issues on our UK cluster and US clusters, resolved it within 30 minutes while people were sleeping at local times to both clusters. We've been with 2 other providers over the 4 years of operation and RS have given us a peace of mind and IT coverage that we've NEVER had until now.And yes, this will be expensive but if you are ordering a large amount of equipment you should be able to get a deal and still have an in-house SysAdmin for less than hiring a full staff in house to cover everything 24x7x365. Oh also, you should think about the following while making your choices: How much downtime is acceptable for your business? I think for most companies, 1 hour is an acceptable amount (once in a while of course). So far our uptime is 100% in the US and UK with clusters with very few single points of failure.
a web site with $10+ million per year in sales shouldn't choose a server management company from a forum as they usually can't take the risk of finding out that the company really suckThere's nothing wrong with getting opinions on this here. Really, this is the best place FOR that kind of thing.
so i guess its better than going and checking your self rather than waiting for the opinion of some ppl that might turn to be wrong
There, you're truly correct. However, there are ways to verify this and check that the company isn't lying on their own webpage. NEVER trust anything that a company puts out on their own page, unless they offer alternative URLs for verification as well.
Now, back to the topic @ hand:
Most of the server management discussed and offered on this forum is for small tasks, such as single servers, for installing software, upgrade kernels, setting up apache etc, ie: for people who have no technical know how when it comes to managing a server.
This may be true for MOST, but I can think of a number of them that actually offer more than just that. You'll want to stay away from the 'cookie cutter' companies though.
A suggestion:
Go through the server management companies listed here (<!-- m --><a class="postlink" href="http://www.webhostingtalk.com/wiki/Companies_providing_server_management">http://www.webhostingtalk.com/wiki/Comp ... management</a><!-- m -->). Find the ones with websites listed and view them.
Get a list of the ones you liked and send them a mail (best bet is to use a 'contact form' if they have one). Give them the situation, and ask them for a price quote.
Even better? Ask for a call back to discover the potential relationship with your company.
Pick one, give them one (just one) of your servers to manage for a brief period (say 30 days). This way you will know if you picked the right one, and how well you work with the company, and they fit your needs.
As for cost? The response will really vary with what you need. Any company that quotes you a specific 'cost' without getting to know your needs personally is only going to cause more problems in the longrun.Greetings Paul:
Depending on the services you need, a management team dealing with mission critical servers and services, offering 15-minute or less response time 24x7x365 et all can run anywhere from $100 to $250 per hour.
Thank you.Agreed that there is nothing wrong with getting information or even hiring support/engineers from here as many people here are private contractors, the same sort of people you find at IBM, EDS, etc. I know for a fact many people here are former employees of those companies..Just perform your due diligence as you would with anything business related. The single most important thing for your project are the people that work on it, so make sure you can -clearly- communicate your needs and expectations prior to speaking with an engineer.Kind Regards,Thanks for the feedback so far, we have also received a number of good PM's. If you are a Server Management provider who has experience with providing management and 24/7/365 monitoring for clusters (preferably geo load balanced), then feel free to send us a PM and we can talk further about our situation in detail.Rates will vary greatly for something like this. A 3-4x range would not be surprising. Also, you are talking about several different services: DB admin, sys admin, network admin. Each group will have specific rate ranges.
I would expect for an operation such as this you would have an internal sysadmin to co-ordinate this effort (perhaps that you?). I don't want to offend by stating the obvious, but you would be surprised how many requests we get at rackAID that simply make no sense from a technical standpoint or are filled with mis-information.
I would also look for a firm that can list specific personnel that will typically handle certain issues.
Also, your SLA's response time requirements will highly impact your costs. There is a huge cost difference between having someone on-call vs. on-demand.
Don't buy more than you need. Some all-inclusive plan may sound great but can often mean you are paying for more than you need.
You may also want to search around for risk calculators. How much downtime can you afford? I wrote a blog post about this some time ago:
<!-- m --><a class="postlink" href="http://www.rackaid.com/resources/rackaid-blog/industry-insights/hosting_costs_getting_it_right/">http://www.rackaid.com/resources/rackai ... _it_right/</a><!-- m -->
The post is geared to smaller operations but some points are the same.
I would say the idea of downtime even extends to your outsourced IT group. What if they go "down" and cannot resolve a problem for you. What are your fall back options?
Lastly, don't overlook your product vendors. If you bought GSLB, firewalls, etc. Then your vendors may have some support plans or recommended providers.
So, in short ... I don't think there is some ball park number. The upper bound of what you can spend should be set by business objectives not by IT objectives. I would put together a plan, put it out there and see what it costs. Then see if this meets your budget. If not, then some of your business criteria may need to be re-adjusted.A site with $10+ Million in sales is all relative. What kind of transaction volume are you actually doing? Are these small average ticket sales or big average ticket sales? $50 average ticket would be less than 600 transactions per day, whereas $25 average ticket would be over 1,000 transactions per day. Is your volume consistent day to day or do you see large spikes in sales during certain periods such as major holidays?How many total servers are your cluster? How many data centers are involved?I wouldn't say that there is a standard rule of thumb in terms of percentage of sales toward development, management and maintenance costs as every organization is different. What you need to weigh are the costs of bringing such expertise in house versus outsourcing it. As others have suggested, you may have a mix of outside support personnel in addition to on staff personnel. Depending on the variables above, you could be allocating anywhere from 2% to 5% of your revenues towards Web site continuity which would translate into $200K to $500K/year. On the flip side, we have customers with multimillion dollar yearly revenue e-commerce sites spending less than 1% of gross sales in terms of maintenance, support, updates, etc.Do you have any idea what you are spending now?I think its far more typical for a site making $10m+ to have everything but hands/eyes support in-house, relying on the local datacenter providers for hands and eyes, and possibly network management alone.. everything else being handled by staff that work for you, and thus have the site as their one and only priority.$10m in sales would warrant hiring your own server management team so you can be 100% sure of what's going on.
If you don't have that team in-house, then you would have it with whatever outsource provider you hire.
What you might want to do is have a chat with a reputable and established company in the industry that specializes in this sort of thing, such as <!-- w --><a class="postlink" href="http://www.rackspace.com">www.rackspace.com</a><!-- w -->.
They can set you up with a dedicated team of server admins to monitor your application and servers 24x7. It will not be cheap.Or do both of what mrzippy suggested... well sorta.I've finally found the balance of both choices for my company (for our current needs) with RackSpace. We use RackSpace's managed hosting for 2 of our 3 global clusters (EU + NA) and have 1 full time Sr Sys Admin on staff to interact with the RS techs. This allows a company to have full control and faster reaction on most things but also have constant monitoring and 24x7x365 hands on the server level.It's paid off numerous times already where RS have caught issues on our UK cluster and US clusters, resolved it within 30 minutes while people were sleeping at local times to both clusters. We've been with 2 other providers over the 4 years of operation and RS have given us a peace of mind and IT coverage that we've NEVER had until now.And yes, this will be expensive but if you are ordering a large amount of equipment you should be able to get a deal and still have an in-house SysAdmin for less than hiring a full staff in house to cover everything 24x7x365. Oh also, you should think about the following while making your choices: How much downtime is acceptable for your business? I think for most companies, 1 hour is an acceptable amount (once in a while of course). So far our uptime is 100% in the US and UK with clusters with very few single points of failure.