Veroxity Wavelengths • Category Archives • Technical
Authors
Take Action
Categories
Archives
Recent Posts
- New Veroxity Whitepaper: VPLS for Any-to-Any Ethernet Connectivity
- Veroxity’s CEO Featured in Investor.com Article
- From Guest Blogger Jeffrey Harrington – The Outside Plant Puzzle: Pre and Post Build Process
- Customer-Centric Culture
- Google’s Plan for Fiber-Optic Networks
- The Consultative Approach to Disaster Recovery Planning: Part Two
- The Consultative Approach to Disaster Recovery Planning: Part One
- Upcoming Webinar: Any to Any Connectivity
- The QoS Question
- Demystifying Cloud Computing
Tag Cloud
- carrier-class network
- continuous data protection
- customer-centric culture
- data center back up
- Data Networking
- disaster recovery planning
- ds3
- Ethernet
- Ethernet-over-SONET services
- Ethernet solutions
- experimental fiber networks
- Fiber-optic networks
- Fiber network expertise
- fiber optic construction
- fiber ring
- Inc. 5000
- Internet connectivity
- MPLS
- MPLS network infrastructure
- oversubscribed network
- post-build phases of construction
- pre-construction coordination
- QoS
- remote back up data centers
- service delivery timelines
- SONET based infrastructure
- SONET services
- t3
- TDM
- telecommunications providers
- VPLS
2010
The Consultative Approach to Disaster Recovery Planning: Part One
From a service provider standpoint, the term disaster recovery is very broad. It has varying levels of importance that are dependent upon organizational goals. It can encompass your network, your databases, your applications, your internet access; again, it’s all about the business need. The scope of a disaster recovery plan changes from one business and industry to another. As a carrier, our goal is to ensure that we have the most survivable network possible so that we can deliver on our Service Level Agreements (SLAs). This is exactly why we have invested in a SONET based infrastructure, developed multiple entrances for lit buildings, and have dual pops in every location so that any one ‘disaster’ can’t bring down customer connectivity.
There are basically three flavors of disaster recovery planning for which clients engage with carriers. They can be provided in conjunction with each other or separately. These plans mean different things to each individual company utilizing them as defined by their business. Each has varying levels of importance based upon business need:
1. Internet Access: A company needs internet access in order to conduct day to day business. However, companies that gain business via the web are looking for diverse routes and more than one carrier. Due to the nature of their business that centers on Internet transactions, they cannot afford to put all of their eggs in one basket.
2. Basic Connectivity: When a business has multiple facilities, connectivity between these locations must be available at all times. From a disaster recovery planning standpoint, a network that will maintain that connectivity to each geographically dispersed location at all times is ideal.
3. Data Center Back Up: It used to be that a company’s primary data center was located at a company headquarters while their secondary backup centers would be remote. We are seeing more and more customers will have a primary, off-site data center in addition to secondary backup locations that are also remote. Connectivity is needed between all of these sites. It is the organization’s business model and goals that determine the need for this connection and at what level. For example, pharmaceutical companies require huge amounts of data storage, so they will keep all information in two or three places. If one of those data centers goes down, this information can be found in the alternative locations. Some companies have applications that are running at all times, i.e. their customers are constantly tracking their online orders or making purchases, etc. If anything happens to either the server connectivity or the servers themselves these companies have an absolute need for another that is standing by, ready to go online.
…
Check in next week for Part Two of The Consultative Approach to Disaster Recovery Planning.
2009
The QoS Question
The buzz about QoS has dwindled since the concept was first introduced but the debate as to whether or not a QoS product is useful seems to be ongoing. Typically deployed on an MPLS network infrastructure, the purpose of QoS is to ensure high priority customer traffic traverses a significantly oversubscribed network.
There is no question that oversubscribed networks need some type of mechanism to ensure no interruption but my doubts about implementing a QoS system are rooted in my view of how a customer might be affected. As a customer, if you are not buying the highest priority level, then your service quality becomes a function of the priorities level purchased by other customers. This leaves you pitted against other customers on the same network where your status can constantly change based on others willingness to purchase an advantage. Increasingly, customers are looking for networks that don’t deploy QoS because they prefer the idea of a level playing field. Their preference is to shape their own traffic before it gets on to the network.
Recently, Veroxity has reached a decision to take a capacity management approach designed to ensure that there is always adequate capacity such that prioritization and classes of service are not required and SLAs are met. Customers that need assistance in shaping their own traffic are offered a managed solution, adjusting their traffic at the edge so that no prioritization is necessary at the core. My belief is that QoS, while positively utilized on some networks, is not something that is a necessity on a VPLS network such as our own.
2009
Diverse Protection or Protected Diversity (Part 1)
This will be my first attempt to clarify a couple telecom terms that are loosely thrown around, but typically misstated and misunderstood. I will try to separate my opinions/descriptions from the definitions. For this discussion, the opposite of Protected is Unprotected. The opposite of Diverse is Collapsed. This first part will discuss Protection. Part 2 will go into diversity.
Equipment Protection – For each Active component in a piece of equipment, a Stand-By component exists to recover in the event of a failure to the Active component.
Critical components (Processors, Matrixes, High-Speed Line Cards, etc.) in a “carrier-class” network should maintain a 1+1 protection scheme. Meaning, for every 1 active component, 1 standby component exists. Less critical components (Low-Speed Line Cards, etc.) in a “carrier-class” network will maintain an n+1 protection scheme. Meaning for a number greater than 1 active components, 1 standby component exists.
Hand-Off Protection – Active and Stand-By Interfaces are utilized to establish connectivity between a Service Provider and End Customer.
Hand-Off Protection is only available for TDM based interfaces. Hand-Off Protection Technologies do not currently exist for Ethernet based services.
Network Protection – A technology which performs pre-defined steps to maintain functionality in the event of a failure.
For SONET services, network protection is ring-based. When an equipment or fiber outage occurs, steps are taken to re-route the traffic over standby network. The standby network is predetermined and reserved. Fail-over time for SONET protection schemes is sub 50 milliseconds. I plan on going into the definitions, advantages, and disadvantages of the available SONET protection schemes in a later blog entry.
For Ethernet or IP services, network protection can come in varying flavors. Data Networks utilize technologies (OSPF, BGP, MPLS, etc.) to re-route traffic in the event of a failure. As the failure occurs, all nodes in the network communicate to determine the best routing for future traffic. Because the re-routing, or convergence, is not predetermined, fail-over time can be from seconds to minutes. Data Network equipment providers are continually striving to develop technologies to minimize the convergence time.
Circuit Protection – For each active-path across a Service Provider’s network, a protect-path exists to recover in the event of a failure.
With traditional SONET based protection-schemes, protection occurs across entire rings, with all traffic on the ring failing-over simultaneously. The major disadvantage of ring-based schemes is that traffic has to hop across rings, frequently collaping into a single location where the rings interconnect. Circuit Protection allows the Service Provider to create two paths from end to end. When traffic enters the network, it is duplicated and sent across both paths. Each path is routed so that no single location is utilized by both paths. When both data streams are received at the last piece of equipment in the network, a selector utilizes the best signal. Circuit Protection is sometimes called Path Protection.
Veroxity utilizes Circuit Protection for routing of all SONET and Ethernet-Over-SONET services. By utilizing Circuit Protection, Veroxity guarantees complete circuit diversity from end to end for each circuit provisioned. Which leads us to our diversity discussion.
Filed under: Technical
Tagged: carrier-class network, Ethernet, Ethernet-over-SONET services, SONET services, TDM
