Data centre

Multiple providers, 10G+ port connections to them with solid networking hardware properly setup, proper fiber with the proper connectiors, route optimizing, using network hardware that works together properly, and if possible at your location try to add peering through an exchange.
 
Multiple ISPs, redundancy network equipments, proper VLAN setup so the customer sharing the same network having problem will not affected other customers reside in the same network.
 
Multiple providers, 10G+ port connections to them with solid networking hardware properly setup, proper fiber with the proper connectiors, route optimizing, using network hardware that works together properly, and if possible at your location try to add peering through an exchange.

Can you explain little bit more about the peering through an exchange ?
 
Can you explain little bit more about the peering through an exchange ?

You can request to have peering exchange with the ISP that near your place so the connectivity between your network and the data center will be connected directly and the hop max will be only 2-3 hops.

Besides that, if there is any internet exchange available in your country such as MYIX, SGIX, HKIX and etc, you may requested to join in and do the direct peering in this IX community as well.
 
You can request to have peering exchange with the ISP that near your place so the connectivity between your network and the data center will be connected directly and the hop max will be only 2-3 hops.

Besides that, if there is any internet exchange available in your country such as MYIX, SGIX, HKIX and etc, you may requested to join in and do the direct peering in this IX community as well.

That looks very interesting. Gives you some good options. Great info :)
 
It depends what kind of context we're talking here.

We have cabinets in various DC's and the big thing that I look at is the number of providers on-net in the facility. If they're on-net, then I don't need to pay darkfiber costs and add unnecessary hops in the network.

Backup power is big one too. Certain regions of the US for example can be prone to certain weather related issues and it's always helpful to know that there's backup power. I remember a few years ago when BurstNET went down because their DIY DC in Scranton was a bodged as you could get and when a storm hit, there was no backup. I think it was soon implemented after that incident.

On-hands staff is a big one too. We don't have people anywhere near most of the facilities we use, so it's very helpful to know that if something goes wrong in the middle of the night, there's someone on-site that can be tapped in to.

There's different types of DC's. Some have their own bandwidth mix that they can offer to customers, others don't and just provide the space & power, then you pull over your own connections from network providers.

NO shortcuts should be taken in this game.
 
A word of caution about number of providers on-net in the facility. I'm aware of one facility that advertised providers on-net that were not there. Don't know if that's the same now. Always verify. Also check to verify that they have level 3 techs on call.
 
A word of caution about number of providers on-net in the facility. I'm aware of one facility that advertised providers on-net that were not there. Don't know if that's the same now. Always verify. Also check to verify that they have level 3 techs on call.

Sensei what do you mean for level 3 support ?
 
Sensei what do you mean for level 3 support ?

A majority of data centers do NOT have fully qualified engineers on-site overnight. In fact, many don't even have a soul on-site overnight.

Companies who operate several facilities like CoreSite, for example, have central monitoring of all their facilities and there are engineers 'on-call' in the event of emergency.

The reason for this? A lot of things can be maintained remotely now too. Switches can be managed remotely, servers can be managed remotely, power control units can be managed remotely. There's no need in many circumstances to have someone.

Some facilities have fully qualified engineers readily available and on-site 24/7/365. Just because they advertise 24/7 support, doesn't mean that someone is on-site.
 
A majority of data centers do NOT have fully qualified engineers on-site overnight. In fact, many don't even have a soul on-site overnight.

Companies who operate several facilities like CoreSite, for example, have central monitoring of all their facilities and there are engineers 'on-call' in the event of emergency.

The reason for this? A lot of things can be maintained remotely now too. Switches can be managed remotely, servers can be managed remotely, power control units can be managed remotely. There's no need in many circumstances to have someone.

Some facilities have fully qualified engineers readily available and on-site 24/7/365. Just because they advertise 24/7 support, doesn't mean that someone is on-site.

Coming from what you said, recently i went to one data center in Sofia, in one of the server rooms there was one person as tech support. His shift wasn't 24h for sure:)
 
Can someone explain little about pci and hipaa compliance ?

Sure.

I have good knowledge of both, more so the latter.

PCI is a set of rules & regulations that must be adhered to in order to maintain standards for the storage and processing of credit card information. You can NOT do so by law without being PCI compliant. That's a basic summary of it).

On the flipside, you have HIPPA which is for the safe storing of medical information. I've done business with many healthcare businesses over the years and HIPPA is a must-have for any data center business wishing to target this industry. There are certain centers who specialise in this industry in particular and employ people to maintain those standards.

It's all about protecting vulnerable information at the end of the day and if you wish to deal with businesses in that realm, then you need to maintain those standards, by law.
 
Sure.

I have good knowledge of both, more so the latter.

PCI is a set of rules & regulations that must be adhered to in order to maintain standards for the storage and processing of credit card information. You can NOT do so by law without being PCI compliant. That's a basic summary of it).

On the flipside, you have HIPPA which is for the safe storing of medical information. I've done business with many healthcare businesses over the years and HIPPA is a must-have for any data center business wishing to target this industry. There are certain centers who specialise in this industry in particular and employ people to maintain those standards.

It's all about protecting vulnerable information at the end of the day and if you wish to deal with businesses in that realm, then you need to maintain those standards, by law.

Thank you for the clear and exact explanation. Now i get it. It has a lot of logic in these regulations for sure. The question is if data centers complies with these rules when storing vulnerable information.
 
Top