Bay Area Network Systems

Building an enterprise cloud

admin — Mon, 07 Nov 2011 19:13:04 +0000

There are plenty of cloud based solutions out there that take on the “enterprise” moniker, but what does it truly mean? I believe that is subjective. You’ll hear phrases like “create a new server instantly with one click” and “we only charge you by the hour for what you need”.. Now, I’m sure there are people out there doing development or other projects that do need to instantly spawn new servers, and this is probably great for them, but is it what you want for a stable enterprise environment?

Our philosophy is stability through planning and implementation. Our hosting experts will build your cloud around your needs taking into consideration several key areas:

Bandwidth: How much do you need? Shared? Dedicated? What general type of content do you plan to be serving up?
Networking: How will your networked VM’s be interconnected? Do you need a DMZ for your app servers? A protected LAN segment for your DB servers? A VPN link to your physical office location?
High Availability: What kind of assurances do you need to make sure your application is always available? Need redundant app servers behind a load balancer? Database cluster? Redundant SAN storage with off site replication for disaster recovery?

Once implemented, we’ll continue to monitor and adjust your CloudStacker environment as needed.

Have any questions? we’d love to hear from you!

Navigating Cloud Storage

admin — Sun, 06 Nov 2011 16:02:36 +0000

It’s easy to go out and pick a hosting solution for your application with the wealth of options out there. But what questions to you need to ask potential hosting companies? For me, hosting an application or an environment always comes down to how it’s stored. It’s the core of your applications availability and survivability. Without a proper storage solution in place, what happens if drives fail? (and they do), what happens if the storage hardware fails? (and it can). Do you have a disaster recovery plan in place in case your provider looses your data?

These are all critical questions.

We have many options available to our clients to fit many different needs and scenarios which we fit into the following categories:

Standard local storage: Great for commodity app servers that house static data. It also is the most cost effective.
High speed local storage: This is often used for high throughput use cases, like transactional database servers. It can offer you the fastest read and most importantly (in most cases) write speeds. This is acheived by using a directly attached RAID 10 array using enterprise grade 10 or 15k SAS disks.
Standard SAN: Using a storage area network gives you the most flexibility. It allows you to use native cloud failover features for high availability. It also gives you a good level of security and redundancy for your data (redundant disks, redundant controllers, etc)
Highly Available SAN: This is where storage availability is taken to the next level. Not only do you have redundant heads, redundant disk shelfs and redundant disk volumes.. but a secondary separate backup SAN that receives file updates from the primary set.

We are also excited to start offering ZFS based SAN units for our customers. With ZFS you get some exciting new features that just wernt available before with previous generation SAN hardware. To name a few:

Automatic snapshot generation: Setup automatic snapshots ever hour (or other interval) to protect your data. Accidentally overwrite or delete a file? Just go back in time into the previous snapshot and recover the file or folder. Snapshots can have a given life to them based on how much you intend to store. By default, this is usually a week but it can be customised for each individual customer.
Write intent log (or ZIL): A write intent log aids in the use of synchronous writes. The client no longer needs to wait for the physical spindles of the disks to finish writing the data.
SSD Read Cache (or L2ARC): We can allocate an SSD read cache in front of a customer’s storage volume so that regularly accessed files are cached which results in huge decreases in read operation times.
Data scrubbing: In previous generation SAN’s, you were not alerted to data corruption until a physical drive reported operation hardware related problems. With ZFS, checksums are stored for all files and the volumes are regularly “scrubbed” to ensure data integrity. That means no silent data corruption. When using mirrored volumes, this means the scrubbing operation can repair the files with the alternate data set.

This is just the beginning when it comes to cloud based storage, if you have any questions, we’d love to hear from you!

Maintaining Your Cloud Environment

admin — Tue, 15 Feb 2011 16:01:18 +0000

In the hosting industry, one question often asked is: “why can’t I just host my own cloud environment on my own hardware?”

While that is certainly an optional route to go. If you crunch the numbers you’ll realise that not only is it not cost effective, you loose out on the majority of the important enterprise features. When you utilise the CloudStacker hosting environment, you get far more than just hosting services. You get an entire support network that can be as involved as needed in your operations.

Here are a few often overlooked items when considering a self hosted environment on client hardware:

Bandwidth: Most people do not factor in the true amount of bandwidth needed during busy usage times for their application. Sure, colocation is cheap when all you need is a few IP’s and shared bandwidth with a small pipe. What happens to that cost when you need dedicated un-metered bandwidth in the 20. 50 and even 100+ megabytes per second range? and what happens when you need even more? will it burst? at what cost?
Hardware support: Who will be on call and at the datacenter when hardware fails? What happens when that person is unavailable or on holiday?
Application support: What happens when your application fails due to issues like DB usage (failed re-indexing, log file growth, transaction performance problems). Do you have an expert on hand that can deal with these issues 24/7/365? (and someone to back him up when he is not there?)
Physical hardware: Are you aware of the investment it takes to get production grade redundant hardware? Firewalls, load balancers, switches, IPS/IDS protection, servers and SAN systems?

Calculate all that together and I think you will realize the considerable savings you can achieve by offloading your hosting and datacenter services to us.

Rotator One

admin — Tue, 15 Feb 2011 15:12:16 +0000

Let us build and deploy your enterprise cloud.

Learn More

Rotator Two

admin — Tue, 15 Feb 2011 15:11:25 +0000

Your data is critical.
Find out how we can ensure it remains safe and available.
Learn More

Rotator Three

admin — Tue, 15 Feb 2011 15:07:28 +0000

Leverage our cloud expertise.
We are experts at moving apps to the cloud.
Learn More